CASE Network Report 54 - Poverty Dynamics in Poland. Selected Quantitative Analyses

Poverty Dynamics in Poland
Selected quantitative analyses
Miriam Beblo
Stanis³awa Golinowska
Charlotte Lauer
Katarzyna Piêtka
Agnieszka Sowa
CASE – Warsaw / ZEW – Mannheim

This research was made possible thanks to the financial support of Volkswagen Foundation.
Review by Irena Topiñska, Phd
Key words: alcohol abuse, education, labour supply, poverty, social exclusion, social transfers,
unemployment
© CASE – Center for Social and Economic Research, Warsaw 2002
Graphic Design: Agnieszka Natalia Bury
DTP: CeDeWu Sp. z o.o.
ISSN 1506-1647, ISBN 83-7178-296-9
Publisher:
CASE – Center for Social and Economic Research
ul.Sienkiewicza 12,00-944 Warsaw,Poland
tel.: (4822) 622 66 27, 828 61 33, fax (4822) 828 60 69
e-mail: case@case.com.pl

Contents
Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1. Poverty in Poland: Causes, Measures and Studies . . . . . . . . . . . . . . . . . . . . 11
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.1. The Causes of Poverty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.1.1. The Communist Legacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.1.2. Transition and Poverty. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.1.3. Challenges of the New Economy Versus Poverty . . . . . . . . . . . . . . . . . . 18
1.2. Defining and Measuring Poverty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.2.1. The Poverty Line Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.2.2. Measures of Poverty Applied in Poland . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.2.3. Research into Poverty. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.3. The Picture of Poverty in the 1990s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
1.3.1. Most Vulnerable Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
1.4. Concluding Remarks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2. Labour Supply Effects of Social Security Transfers . . . . . . . . . . . . . . . . . . . . 37
Katarzyna Piêtka
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.1. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.1.1. Review of Social Policy in Poland During the Transition . . . . . . . . . . . . . . 38
2.1.2. Social Expenditures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.1.3. Individual Transfers to Households . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.1.4. Unemployment in Poland . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.2. Empirical Study. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.2.1. Goals of the Investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.2.2. Statistical Analyses of Work-Related Characteristics
among Transferees and Others . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.2.3. Econometric Analysis of the Relationship between Transfers
and Labour Supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.3. Concluding Remarks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3

3. Alcohol Abuse and Poverty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Agnieszka Sowa
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.1. Alcohol Abuse as a Social Problem in Poland. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.1.1. Alcohol Abuse, Alcohol Dependency and Alcohol Related Problems
– Towards a Definition of the Phenomenon . . . . . . . . . . . . . . . . . . . . . . 70
3.1.2. The Individual and Social Consequences of Alcohol Consumption . . . . . . 72
3.2. Empirical Study. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.2.1. Research Hypothesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.2.2. Data Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.2.3. Methodology of the Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.2.4. Characteristics of the Studied Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.2.5. Factors of Reporting Alcohol-related Problems by Social Assistance
Beneficiaries – Outcomes of the Research . . . . . . . . . . . . . . . . . . . . . . . 84
3.3. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4. Family Background and Children's Educational Attainment During
Transition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Miriam Beblo and Charlotte Lauer
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.1. Poverty During Transition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.2. Children's Education as a Mechanism of Poverty Transmission. . . . . . . . . . . . . . . . 94
4.3. An Empirical Analysis of Children's Educational Outcomes . . . . . . . . . . . . . . . . . . 97
4.3.1. The Polish Education System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.3.2. Enrolments and Educational Attainment During Transition. . . . . . . . . . . . 99
4.3.3. Determinants of Educational Attainment . . . . . . . . . . . . . . . . . . . . . . . . 108
4.4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
4
CASE Reports No. 54 – Poverty Dynamics in Poland

Contributors
Miriam Beblo studied Economics at Technische Universität Berlin (Germany) and
University of British Columbia (Canada). After being a research/teaching assistant and
completing her doctoral thesis at Freie Universität Berlin (Germany), she joined ZEW in
2000 where she is working at the department of Labour Economics, Human Resources and
Social Policy. Her research focus is on the labour market effects of social policy, particularly
family policy measures, as well as on the analysis of income distribution and wage structures.
Stanis³awa Golinowska, Professor of Economics, studied Economics at Warsaw
University and at Mannheim University as a scholarship-holder of the Humboldt Foundation.
From 1991-1997 she was a director of the key research institute in the fields of the labour
market and social affairs – IPiSS in Warsaw. She is associated with the CASE Foundation as
a founder and vice chairman of the Foundation council. Now she is also teaching social policy,
social insurance and health economics at the Jagiellonian University in Cracow. Author of
numerous articles and books on social aspects of economics and social policy reforms.
Charlotte Lauer studied Economics and Business at the Ecole Supérieure des Sciences
Economiques et Commerciales (ESSEC) in Paris. She then completed a Master of European
Economics at the European Institute at the University of Saarland. She joined the ZEW's
department of „Labour Economics, Human Resources and Social Policy” as a research fellow
in September 1998. Her current research interests concern issues related to the economics
of education and labour market policy. Charlotte Lauer is an associate researcher at the
institute ERMES at the university Panthéon-Assas in Paris (Paris II).
Katarzyna Piêtka studied Economics at Warsaw University. She is an economist at
CASE Foundation in Warsaw since 1994. Involved in research projects as well as advisory
programs in Central and East Europe (Ukraine, Belarussia); currently also a co-editor of the
macroeconomic quarterly Polish Economic Outlook. Her research topics focus on social
policy, the labour market and macroeconomics. The author of several studies about the
pension system, transfers to the disabled, and the labour market for the elderly.
Agnieszka Sowa studied Sociology at Warsaw University with specialisation in public
policy (Pittsburg University). Joined CASE Fundation in 2001. Involved in Research projects
on socisl policy and social exclusion, labour market and health care. Currently participating
in ILO course in social protection financing at the Maastricht University.
5

Abbreviations
CASE – Center for Social and Economic Research
CBOS – Center for Public Opinion Studies
GUS – Central Statistical Office
HBS – the household budget survey
IFiS PAN – Institute of Philosophy and Sociology at Polish Academy of Science
IPiSS – Institute of Labour and Social Studies
IPS UW – Institute of Social Policy of Warsaw University
ISP – Institute of Public Affairs
KRUS – Pension Fund of the Agricultural Social Insurance Fund
LFS – Labour Force Survey
MPiPS – Ministry of Labour and Social Policy
PARPA – State Agency for Prevention of Alcohol-Related Problems
POMOST – data base of the social assistance centers
SAC – Social Assistance Centres
ZEW – Centre for European Economic Research
ZUS – Employee Social Security System
6

The present report summarises the outcome of a research project carried out jointly by
researchers of the Polish Center for Social and Economic Research Foundation (CASE) and
the German Centre for European Economic Research (ZEW) and funded by the Volkswagen
foundation. The objective of this project is to analyse the mechanisms at work in the rise and
persistence of poverty during transition in Poland, as well as its consequences for selected
groups of the population.
The transition process from a centralised to a market economy in Poland has been
accompanied by an unprecedented increase in poverty and a deepening of inequality across
households – not only in terms of income but also in terms of socio-economic status.
Although a small number of studies describing the economic situation of the poor in Poland
have been undertaken, our understanding of the mechanisms that make poverty persist in
the household context is considerably limited. The interaction of a number of factors may
for example, result in individuals being trapped in a vicious circle of poverty. Low household
income may lead to social exclusion and family distress, which is likely to have far-reaching
consequences for all household members. Social exclusion may contribute to foster
alcoholism, impede the human capital investment in children, and thus jeopardise the socio-economic
situation of the next generation. Socially excluded people experience severe
difficulties in finding re-employment. Social transfers might even worsen the situation by
providing a disincentive to seek work.
We need to understand the causes underlying the developments in social and economic
hardship of Polish families during the course of the transition process. The introductory
chapter therefore offers a general look at the picture of poverty in Poland; trends and new
research results are described. In order to improve our understanding of the causes of social
exclusion and to contribute to filling the gap in research we do not, however, restrict our
attention solely to the analysis of the extent and nature of poverty in general but rather focus
our analysis on issues that have been somewhat overlooked. This project contributes to the
literature by investigating empirically different dimensions of the poverty debate in Poland –
7
Introduction

ranging from social exclusion through the relationship between transfers and labour supply
to the transmission of poverty across generations. The empirical analyses are carried out on
the basis of individual and household histories which are observed in the Polish Labour Force
Survey and of administrative data on social assistance beneficiaries.
In the first empirical study we examine the relationship between social transfers and
labour supply: although poor people may apply for social benefits, the benefit entitlement
may result in a poverty trap by providing disincentives to acquire work. Moreover, long spells
of poverty without employment are likely to decrease a person's chances in finding a new
employment position. The extent to which individuals react to these disincentives and/or are
excluded from the labour market after long periods of non-employment has been estimated
using the Polish Labour Force Survey – for the periods 1997 and 2000. Applying a logit
regression, we test how different transfers and other kinds of non-earned income – given
individual, family and regional characteristics – affect the decision to look for a job or not.
The results obtained indicate that the propensity to search for employment is highly
differentiated between different groups of transferees. The more stable transfers are
financially, the lower is the labour market activity of their recipients. Unemployed individuals
who are entitled to benefits are less eager to work than registered unemployed maintained
by others. The longer a person is off work, the less willing he/she is to search for a job. These
findings, however, have to be interpreted with caution, since: i) only proxy methodologies
are used in the study, and ii) because of the problems involved in establishing the direction
of causality. In addition to receiving transfers, a number of individual characteristics are also
found to play an important role in the attitude to work. More specifically, a lower number of
children but higher number of adults in the family, the main household income being derived
from transfers, living in big cities, being male and unmarried are all positively related to job-search
intensity.
In the second study, we look at an essential factor influencing social exclusion in Poland:
alcoholism. GUS reports a strong increase in alcohol consumption in Poland since the beginning
of the 1990s. Public opinion considers alcohol dependence as one of the most important social
problems influencing household's economic performance. The problem of how family and
economic situation correlates with alcohol dependence among the poor population is
therefore investigated. If being poor and alcohol dependent is correlated with certain types of
labour market activity (or lack of such activity) and some family characteristics, then describing
these relationships is an important policy input. However, we are aware that it is impossible to
state causal relations between alcohol abuse and the economic or social situation of a
household as these problems are inseparable/interdependent. To target the poor population of
alcohol dependants we concentrated our analysis on social assistance recipients, using
administrative data from the sample of social assistance centres. The data provides individual
and household information for every social assistance recipient. The analysis was conducted on
8

the group of recipients who stated alcohol problems together with their poor economic
performance and therefore turned to a social assistance centre for help. The outcome of a logit
model indicates that reporting alcohol abuse among social assistance recipients is significantly
related to gender and labour market activity, family composition and living in rural areas. It is
important to note, however, that this research does not cover the whole complex problem of
alcohol abuse in Poland. It shows some important relations between alcohol abuse,
employability and the household situation that in consequence might lead to household
deprivation and social exclusion.
Finally, we examine the extent of intergenerational transmission of poverty through the
education of children: The starting point of this chapter is the hypothesis that children are
most likely to bear the consequences of their parents being caught in a poverty trap. We
therefore estimate the effects of economic hardship of the parents on the situation of the
children, in particular with respect to their educational attainment. If parental poverty has a
large impact on the educational prospects of the children, poverty is likely to be passed on
over the generations. If this happens to a large extent, the transition process may exhibit a
detrimental dynamic by which some children are disadvantaged under the new economic
system from the very beginning of their lives. Societal disruptions may follow in the medium
run. Using the Polish Labour Force Survey we analyse the intergenerational transmission of
poverty from Polish parents to their children through children's educational attainment
during the transition process of the 1990s. The relationship between family background and
education is investigated using an ordered probit model of educational attainment. The
results illustrate that children's education is strongly related to household structure, parents'
education, city size, and region of residence. Household income and the parents' labour
market situation have only a weak, though significant, effect on children's education. We
conclude that, if poverty transmission takes place across generations, this seems to be
primarily caused by the inheritance of human capital rather than by pure wealth effects. This
link has remained comparatively stable throughout the transition process.
The three issues outlined above may be interrelated through a possible line of causation
running from long-term unemployment and social transfers dependence, through social
exclusion and poverty – often accompanied by alcohol problems and family dysfunction, to
the transmission of poverty across generations. If benefit levels are very low, their
contribution to the alleviation of poverty is minimal. Children of deprived families have
relatively worse educational prospects than those with better-off parents. This leaves them
with a comparatively low earnings potential in the future. As a consequence, these children
are more likely themselves to have families burdened by economic hardship, and so the
vicious circle closes. By providing the links between our three studies we aim to highlight a
number of issues of particular importance to the poverty debate in Poland today that may
help to improve the effectiveness of the Polish social policy.
9
Introduction

The remaining of this report is structured as follows: Part 1 provides an overview of the
prevailing problem of poverty in Poland. The labour supply effects of social transfers are
studied in Part 2. After an investigation of alcohol abuse among social beneficiaries in Part 3,
Part 4 tackles the transmission of poverty from parents to children through the educational
attainment of Polish children.
10

Introduction
Running parallel to the impressive development of Central and Eastern Europe over the
last decade has been an unprecedented increase in poverty. Despite the notable
achievements in developing a market economy and fostering economic growth in the region,
the increases in inequalities and labour market problems that have also arisen were not
expected on such a large scale and for such a prolonged duration. "The presumption was that
growth would come quickly as countries move forward with the transition process, and that,
with good mobility and high level of education, it would reduce the incidence of poverty
rapidly. Poverty was believed to be largely transitory in nature, and best addressed through
the provision of adequate safety nets" (World Bank 2000, p.V).
The problem of poverty in Poland is especially important as it is connected both with old
– and at the same time new – problems of development. New problems have stemmed from
the deep restructuring of the economy, which brought with it an often painful deterioration
in the labour market. Old problems relate to the fact that there was already a significant
number of hidden unemployed and socially excluded members of Polish society.
This chapter offers a general overview of poverty in Poland: trends, measurements and
research results, at times in comparison with other countries in the region.
1.1. The Causes of Poverty
Even the briefest of analyses of the ways in which Polish society has developed in recent
decades tends to indicate that the growth of poverty is connected with two main processes:
the first is the opening up of the agricultural sector to strong market competition, with the
11
Part 1
Poverty in Poland: Causes, Measures
and Studies

subsequent collapse of many small farmers whose existence had been based almost
exclusively on the subsistence (natural) economy.
The small farms sector is no longer able to support the subsistence needs of the rural
population. Added to this, widespread bankruptcies of large state-run farms led to large
numbers of redundancies in other entities that had been economically interdependent on
them. As a result the situation in the Polish countryside deteriorated dramatically and the
incidence and degree of poverty in rural areas has become a very serious social and
economic problem.
The second process has been connected with changes in industrialisation and
urbanisation. Poland's traditional industries have been undergoing massive restructuring for
some years. One of the key social questions associated with these large-scale and
fundamental changes in the industrial structure of the Polish economy has been the extent
to which employees – the most affected by the changes, have been willing and able to adapt
to the new circumstances. Failure to do so in many cases has become a further additional
source of poverty, this time largely in urban areas. Such industrial changes have to some
extent been ameliorated by cross-party support for social programs to ease employees out
of traditional workplaces. This has meant that the poverty of these groups has not always
been directly connected with redundancy, but with later adaptational difficulties.
1.1.1. The Communist Legacy
Despite great efforts to become an industrial powerhouse via rapid industrialisation and
intensive modernisation programmes undertaken during the communist period, Poland failed
to catch up on the industrial front with the more advanced economies of Western Europe.
Investment plans were poorly worked out and poorly executed, and there was a consistent –
and systemic – lack of capital, despite the low level of consumption and large-scale income
redistribution. In these conditions incentives to modernise were low. In addition, the political
weakness of the authorities further limited possibilities of achieving such aims.
Such "half-baked modernisation" in turn opened up and accentuated three main problem
areas: (1) disparities between the living condition in rural and urban areas, (2) strong regional
disparities, and (3) low education level of the workforce and low adaptational abilities of
certain groups of the population.
Neglect of rural areas and agriculture
During the three decades of accelerated industrialisation (1949-1978) Poland's private
agricultural economy was a key source of labour for industrial growth. The authorities'
exploitative attitude to agriculture began to change somewhat in the mid-1960s, though a
12

significant improvement in the rural population's standard of living did not appear until the
1970s, when the authorities significantly reversed their policy towards agriculture and rural
areas. Changes began with the removal of key regulatory barriers (primarily the removal of
the obligations on farmers to supply agriculture products). In the second half of the 1970s a
wide range of investments into the agriculture sector were conducted. The prices of
agricultural products, primarily meat products, sold to the state were increased, alongside
the simultaneous introduction of advantageous credit policies for farmers, subsidised feed
and synthetic fertiliser prices.
In 1977, social insurance for farmers was introduced on similar bases to those for non-agriculture
employees; up to that point farmers had been entitled to pensions in return for
bequeathing their land to the state.
The change of policy towards agriculture, including private farming, led to a situation in
the 1980s in which farmers' incomes had risen to match those of non-agricultural employees,
and by the end of the decade they had surpassed them.
But this change did not bring any significant improvements in the land-use structure or
major improvements in the technical infrastructure in the countryside (despite private
initiatives undertaken to this end, such as the Church Fund, and later, the Village Support Fund).
Regional disparities
Throughout the post-war period the state failed to tackle the underdevelopment of some
regions in Poland (especially in the former Kongresówka areas in eastern and central parts of
the country). The considerable regional disparities that grew in the past have not diminished
in the transformation period since 1989.
Educational backwardness
One of the main legacies of communist Poland's educational policy is the dominance of
basic vocational training. The accelerated industrialisation of the country over three decades
(1949-1978), with an exceptionally intense phase in the first half of the 1950s, demanded a
massive increase in the number of highly qualified workers. The vocational secondary schools
and its mass growth were seen as the best way to meet this demand. Today, most middle-aged
working people only attained basic vocational education (over 50% – Cichomski op.cit.).
The financial situation of this group was relatively good and many so-called 'labour elite' and
future trade unionists originated from this group, many even ascending to political office in the
1980s and 1990s. The majority of this group became owners of the small- and medium-sized
private companies that sprouted en masse in the period 1989-1992 (Domañski 1998).
Today, however, the politicisation of this group, its only partial acceptance of market rules
(especially related to efficiency), as well as lack of willingness to adapt to changing situations
on the labour market are today key barriers to development.
13
1. Poverty in Poland: Causes, Measures and Studies

1.1.2. Transition and Poverty
Studies on poverty around the world have shown that poverty in transition countries is
somewhat different from poverty at other places. The first feature of poverty in transition
countries is that those who are poor tend to be what we can term 'the new poor'. In the
communist period all had guaranteed employment and regular pensions from the state after
retirement. While average incomes were low, income inequalities were insignificant as
people's predicaments were similar. Under communism, the greatest concern was access to
goods and services. The central planned economy created shortages and people were unable
to satisfy their basic needs. There were of course some people living in what we may call a
'culture of poverty', but such groups were rather marginal and the state was able to deal with
them in various restrictive ways, for example via institutionalisation, in one form or another.
The introduction of the market economy opened up a huge consumer market and greater
opportunities to achieve higher incomes, though not for all. The unemployed, small farmers,
adults and children from dysfunctional families, for example, had fewer chances of succeeding
and difficulties in adjusting to the market system caused many of them to fall into poverty.
The transition from the centrally planned to the market economy accelerated changes in
agriculture as well as industrial restructuring. Poland is following developmental trends that
have been prevalent in more advanced economies over the last few decades. The changes in
Poland, however, have been exceptionally intense on a regional basis due to the fact that the
country's starting point was that much lower and the political incentives introduced in Poland
that much stronger. Transformation in the agriculture had been held off for a very long time.
Changes in rural areas
The period of transition to the market economy, together with the liberalisation of
foreign trade, brought considerable problems for Polish agriculture and the rural population.
The profitability of agricultural production decreased radically – though the "insatiable
domestic food market" had already before this become less robust on the demand side. The
subsequent improvement in farmers' material situation was primarily associated with the
transition from the agricultural to non-agricultural economy, though this had become far
more difficult than in earlier years. Internal migration from the countryside to the city and
the two-occupational population (farmer-worker) dropped off.
At the end of the 1980s the real income of households involved in agriculture was
growing by up to 20% a year. In the 1990s farmers' average income decreased in real terms
by approx. 38%. The earnings of employees exceeded income in rural households by
approx. 30% on average. These incomes stayed more or less stable for the next few years,
dipping again in the second half of the 1990s. At the same time the real incomes of other
social and economic groups, including the employed and self-employed, grew steadily after
14

1994. As a result, the ratio of rural and other household average income to the average
income of the employed households has changed (Table 1.1).
Table 1.1. Ratio of rural and other household average income to the average income of the employed
households (the average income of the employed households = 1.00)
The decrease in the average agricultural income – both in absolute terms and in relation
to other groups of households – was a result of systemic changes and the development of
market mechanisms in the economy. The abolition of budget subsidies to food production –
intensified in 1989-1990 – was accompanied by liberalisation of prices and margins as well as
the appearance of imported food on the market. For the first time in the post-war period
there were meat and dairy product surpluses in Poland.
Changes in the labour market was the second key factor lowering the income of the rural
population. In the initial stage of transformation (up to 1993), the number of people working
in agriculture decreased. This was due, amongst other things, to increased incentives for
older farmers to retire. However, since 1994 the number of the people working in
agriculture has been growing, largely due to structural and institutional changes on the labour
market outside agriculture. The absorption of labour surpluses by the non-agricultural
economy was very slow and, at the same time, eligibility for unemployment benefits was
restricted. Accordingly, farms became kind of "storerooms" for people who could not find
jobs outside agriculture and who were not entitled to unemployment benefit. As a result,
relatively high hidden unemployment in the individual farming sector reached, according to
some estimates, 900,000 people (Kryñska 1999).
At the same time, mainly due to the collapse of state farms, there was also high official
unemployment among that section of the rural population not possessing any land of its own
and not involved in individual farming.
As a consequence, poverty appeared in the countryside, though of a somewhat
different character to the sense in urban areas in that prevailing that the existence-threatening
factors are far less visible; one may call it 'cultural poverty' to a far greater
extent than the label applies to urban poverty. Though it is hard today to imagine that
people possessing even a small piece of land may suffer hunger, the phenomenon of
15
Economic
group of
households
1989 1993 1994 1995 1996 1997 1998 1999 2000
Rural 1.16 0.89 0.87 0.94 0.87 0.83 0.76 0.69 0.69
Self-employed n.a. 1.24 1.27 1.28 1.23 1.23 1.22 1.21 1.21
Pensioners 0.73 1.05 1.05 1.04 1.02 1.04 0.99 1.01 0.94
Source: UNDP 2000 on the basis of data from the GUS.

limiting to the subsistence economy is relatively common. Cash in many places has
become a deficit commodity and is used to cover the costs of mobility and non-foodstuffs.
Poverty in the countryside has become a wider phenomenon since 1997. Before this,
poverty was concentrated mainly in the former state farms. Today it also affects individual
farmers' households to a considerable extent.
Restructuring
Industry has been affected by equally deep-seated changes. The restructuring of
traditional sectors, as part of the so-called 'post-Ford transformation' (Mingione, 1996), is
accompanied by necessary ownership changes, which, however, makes the process more
difficult to co-ordinate and adapt socially. Changes in the industrial structure are often
analysed in terms of the ways and degrees to which they act to create and destroy jobs
(World Bank 2001).
Restructuring of the economy was intensified in the period 1997-2000. One prominent
feature of the process was its asymmetry, with the predominance of job losses over job
creation. In the period 1998-2000 the number of employed people decreased by approx.
360,000 (estimation – Witkowski 2001).
New jobs have been created mainly in the private sector and in services, though in recent
years both have slowed significantly. In the period 1998-2000 over 80% fewer jobs were
created than in 1994-1997 – the period of their most dynamic growth. The decrease in the
number of new jobs in services was even bigger – exceeding 90%. These trends can be
explained by a lower rate of job creation in the private sector overall as well as estructuring
of the public sector (including social services) which even led to the drop in employment.
Restructuring resulted in labour productivity growth. In industry, where the decrease in
the number of employees was greatest (over 550,000 in 1998-2000), labour productivity
grew by 14% in 2000 alone (Witkowski 2001). Growth in industrial labour productivity
reflects a modernising economy and should presage higher efficiency in the economy as a
whole. However, without job creation in other sectors to compensate for the job decline
in industry, such productivity growth has certain negative effects on the labour market, in
particular a dramatic growth in unemployment.
Institutional changes
The transformation towards the market economy has brought with it key institutional
changes. New regulations, commercial values, growth in efficiency requirements,
individualism against higher living standards in small communities all came too rapidly for
many groups and did not leave enough time for necessary adjustments.
Social policy institutions could play a key role in facilitating transformational
adjustments. However, they have been used merely as institutions to bear the brunt of the
16

state's avoidance of the need to adjust to new realities. The old system was based on
relatively generous system of social benefits for people leaving the labour market (pre-retirement
pensions, disability pensions, pre-retirement allowances and benefits).
Inequalities
It has been empirically validated in many case studies that income inequality is a major
cause of poverty in countries with weak economic growth. If inequality grows at times of
weak economic growth and there is no significant income redistribution or income
redistribution may perversely deepen inequalities, then poverty increases. This kind of
connection between inequality and poverty has visibly occurred in transition countries
(Klugman, Micklewright, Redmond 2002) (Table 1.2).
Table 1.2. Gini coefficient for income per capita and for earnings in the post-communist countries
Income disparities between rich and poor in Poland increased substantially during the
1980s and 1990s. The average Gini coefficient of disposable income rose from 28 in the end
of the 1980s and the first half of the 1990s to 33 for the period 1996-1999 (World Bank
2000). This value places Poland, after only several years of transition, at a level near the
OECD (Milanovic 1998).
17
Region/country 1987-1990 1993-1995 1996-1999
Central E. Europe
Czech Republic 0.19 (0.20*) 0.23 (0.26*) 0.25 (0.25*)
Hungary 0.26 0.29 (0.32*) 0.25 (0.35*)
Poland 0.28 (0.20*) 0.28 (0.29*) 0.33 (0.30*)
South E. Europe
Romania 0.23 (0.15*) 0.29 (0.29*) 0.30 (0.37*)
Bulgaria 0.23 (0.21*) 0.38 (0.29*) -
Baltics
Lithuania 0.23 (0.26*) 0.34 (0.37*) 0.34 (0.37*)
Latvia 0.24 0.31 (0.34*) 0.33
Estonia 0.24 0.35 0.37 (0.38*)
CIS
Russia 0.26 0.46 (0.48*) 0.47
Ukraine 0.24 0.36 (0.41*) 0.42*
Note: * Earnings.
Source: Data Bank from Innocenti Research Centre Unicef (2001) and from the World Bank (2000).

1.1.3. Challenges of the New Economy Versus Poverty
Globalisation, work and poverty
The problem of job creation at the present stage of economic development is far more
complicated than ever before. The basic, and clearly identified, global trend in the area of
work is towards a steady fall in employment.
This can be explained in various ways. The process of economic globalisation (primarily
the liberalisation of capital flows) is associated, for some, with both a remarkable leap in
efficiency but also concurrent large reductions in traditional employment. Others suggest
that social policy merely complicates institutional adjustments which would in themselves
help create jobs.
Independent of the arguments that seek to explain falls in employment, one cannot fail to
observe that the "good" job – in its traditional "wrapping" (in other words workplace in its
traditional infrastructure marked by labour laws and industry relations) – has become
increasingly rare.
Table 1.3. Global employment rates
The global trend of falling employment in the poorer worldwide regions coexists in
transition countries with systemic reforms and accelerated restructuring of the economy –
factors making the labour market process additionally difficult. Poland is one of those countries
where additional tensions on the labour market have a demographic source – relatively large
inflows of labour onto the market are expected for the next two decades at least.
The problem of work is therefore of huge strategic importance, and not only a temporary
difficulty at the present stage of transformation.
Since about the 1980s most new evidence from household surveys suggests that
unemployment is playing an increasingly important role as the main cause of poverty,
overtaking low incomes (Lipton 1995), primarily across the industrialised countries. In CEE
countries unemployment appears to have even graver consequences than in EU countries
(Klugman, Micklewright, Redmond 2002).
18
World regions/countries 1990 1994 1997 1999 2000
EU 61.5 - 60.7 62.1 63.1
USA 72.2 - 73.5 - 74.1
Latin America 58.2 58.0 57.8 57.0 -
CEE 70.6 63.5 60.7 63.9 56.3
Poland 83.0 60.8 58.3 59.7 56.4
Sources: World Bank 2001, ECLAC 2001.

In the context of labour market difficulties groups most prone to unemployment are
especially vulnerable. Young people are a case in point (Table 1.4). People with poor
qualifications are a separate matter. Raising qualifications and improving readiness for work
is one of most difficult tasks facing social policy. These groups are most threatened by
poverty and social exclusion.
Table 1.4. Unemployment among vulnerable groups
Increasing skill levels
Developing higher skill levels is one of the most visible challenges for the future, with the
knowledge-based economy at the forefront of any future developed economy. Countries
that choose not to take up the challenge will be pushed to the peripheries of progress.
This is why improving education in Poland demands answers to many questions,
initially in education policy. In the face of a huge increase in the number of young people
education policy sought at first to substantially increase the non-public sector. But the
predominance of cheap mass solutions revealed deep structural deficits in the education
system in Poland, in particular a mismatch between strategic policy directions and labour
market demands for skilled and professional employees. Quality deficits also became
increasingly visible. Mass education was undertaken by teachers often with low skills via
mostly or exclusively oral-based teaching methods (too little practical work and training).
In addition to this, the education process was weakly controlled due to the
decentralisation of the schooling system.
19
Groups of
different
vulnerability
1993
November
1996
November
1997
August
1999
February
2000
1st quarter
2001
1st quarter
On average 14.9 11.5 10.7 12.5 16.7 18.2
Young people
(15-24)
31.6 26.2 23.5 28.5 37.9 41.2
Female 16.9 13.4 13.5 13.5 18.5 19.8
Persons at
8.5 6.5 6.5 6.3 10.6 11.9
immobile age
(over 45 years)
Persons with
poor
qualifications
(basic and lower
education)
15.0 12.9 12.9 11.7 22.1 22.9
In towns 16.9 12.0 11.5 12.9 17.1 19.2
Source: GUS; LFS.

The low quality of the education process in general and large differences in education
results in the latest period are revealed in comparative OECD studies from 2000, known as
PISA (Programme for International Student Assessment)1. PISA research consists of checking
the competence of the 15-year old pupils. PISA measures the basic skills required in modern
societies to lead economically and personally satisfying lives, and also to participate actively
in social and political life.
While the results for Polish pupils were very poor, some Western European countries
where the populations are more heterogenic, that is multicultural, with a large group of
immigrants or with ethnic differentiation (J. Allmendinger, S. Leibfried 2002), had similar
results2. In contrast, Poland, a country with a significantly homogenic population from the
ethnic point of view, has an education system that is rather unequal in quality terms. The low
quality of the education on offer especially affects those with difficult access to education,
that is, young people from the countryside and small towns.
Many of these young people have little of external – that is parental or wider family –
support and this is another factor influencing inequality of opportunity in education. Results
from research in this area reveal that Polish society during the transition has tended to retain
a strong degree of conservatism and societal cohesion. Most children tend to reproduce the
education patterns of their parents. Parents with basic education tend to send their children
to the most basic vocational schools. Parents with higher education likewise tend to seek to
ensure that their children go to university – 70% of them at least make efforts to ensure that
their children attain secondary level education (on the basis of 8 years (1992-2000) data from
the LFS – see Part 4).
The transfer of education levels across generations creates at the same time a mechanism
for reproducing poverty. This requires intervention, both to counteract the process itself and
to tackle the underlying causes of poverty and social exclusion.
Inequalities
Wide income differentials reflect the existence of considerable poverty in all societies.
This derives from the common usage of measures based on the concept of 'relative poverty',
which highlights the gap between the lowest income groups and the average.
Income differential trends in EU countries tend to be split north/south, with the north
much lower than the south, with the exception of the United Kingdom, where inequalities
have been traditionally larger. The rate of poverty in countries with greater income
differentials tends to be higher. In southern countries – the poorest countries in the EU –
relative poverty tends to be the highest (Table 1.5).
20
1 PISA studies are cyclical.
2 The results for Poland have been available for several years, though earlier poor results were associated with the
previous, communist, educational system.

Table 1.5. Diversity of income in the European Union, 1995
1.2. Defining and Measuring Poverty
In the literature there are many approaches to defining and identifying poverty. Some
authors refer to recipients of social assistance as the truly poor because the public support,
though not deprivational in itself, establishes poverty as a social phenomenon of exclusion.
From this point of view, the poor can be identified in two different ways (Mingione 1996).
On the one hand, the poor can be identified as these who apply to institutions for social
assistance. On the other hand, the poor can be defined as those whose living conditions,
consumption or earnings, are worse/lower from an established standard and adopted as desired
(Mingione, 1996). In the first case only recipients of social assistance are considered poor. In the
second case – all those living below the adopted standard.
The first approach, however, can be effectively applied only if the social assistance
institutions do not change their ranges of eligibility and benefits. This would be possible only
in countries with very stable institutional structures in this field.
The poor as recipients of the social assistance programmes
According to the first method, only individuals assisted by specific social assistance programmes
are considered poor. Identification of the poor with social assistance has some strong limitations,
21
Countries Gini coefficient
for income
Change in Gini
coefficients during
the period of time
(in %)
Rate of poverty*
below 50% of
national income
(in %)
Rate of poverty
below 50% of
income for the EU
(in %)
Denmark 0.243 -4.9 (1983-1994) 6.5 4.4
Finland 0.176 9.1 (1986-1995) 4.6 5.8
Sweden 0.242 -1.0 (1975-1995) 5.5 5.9
Norway 0.213 9.4 (19986-1995) 5.0 5.1
Belgium 0.251 2.3 (1983– 1995) 12.6 8.7
Germany 0.251 6.4 (1984-1994) 10.9 7.6
France 0.303 -1.7 (1979-1990) 13.4 10.5
Ireland 0.255 20.0 28.3
Luxemburg 0.298 12.9 1.3
Holland 0.295 11.4 9.2
United
0.357 20.6 14.8
Kingdom
Greece 0.354 21.8 42.0
Spain 0.354 19.1 31.7
Italy 0.348 12,7 (1984-1993) 17.9 23.8
Portugal 0.473 26.8 47.5
Note:* poverty line – 50% of median income calculated for one consumption unit according to the OECD equivalence scale
Source: Vogel, 1998 on the basis of ESSPROS data.

which is why it is used less often than the first approach to poverty. The main problem of using this
method is connected with the obvious difficulty of comparisons over time due to reforms in social
assistance systems and also of cross country comparisons due to highly diversified conditions of the
national social assistance provisions. One very interesting study into poverty in Germany defined
in terms of social assistance (Leisering and Leibfried 1999) claims that it was possible only due to
the relative stability of the country's social assistance system (1961 – introduction of Sozial Hilfe
Gesetz) which made possible a longitudinal cohort survey into social assistance claims.
Dynamic approaches to poverty research (during the life course) have a long tradition.
They have been developed on the basis of two opposite hypotheses. The first, developed by
Benjamin S. Rowntree (a pioneer of empirical research into poverty), stated that poverty
tends to be a temporary phenomenon, with people typically not poor throughout their lives
but only during certain stages: when they have dependent families or their earning power is
limited, for instance, or in the later stages of life (Rowntree 1901).
The second hypothesis assumes that poverty is a chronic and hereditary phenomenon.
Some groups of people live permanently in a culture of poverty. This approach comes from
an American study on chronic poverty among immigrants (Lewis 1959).
1.2.1. The Poverty Line Approach
The poverty line approach is based on establishing a criterion separating the poor from
the non-poor. There are many alternative proposals and discussions on the best wayto do
this, however, and as a consequence different poverty lines have been applied. This does not
help in fully grasping the problem. Politicians and society alike tend to want simple answers
to the questions: is there poverty and how people suffer from it?
The most widely used poverty line approach identifies the poor as belonging to
households with budgets below (officially or statistically) what is defined as 'required income
level' (the monetary poverty measure).
This 'required income level' is most frequently defined either as household consumption
or household income. Much has been written on the relative merits of each measure
(Atkinson, 1998), but no consensus has yet emerge as to which should be applied, though
consumption is more often considered the more appropriated measure. In Polish statistical
analyses expenditure is used as the key indicator of consumption, more than of income.
Another problem is how to set the poverty line: via detailed analysis of consumption and
household needs (the basket method), or by classification of people living at the lowest income
level as poor? In the second case a question appears: how to classify lowest income – as 40%,
50% or 60% of the average or rather against the median? Poverty defined as the proportion of
a given category relative to average income (expenditure) is called 'relative'.
22

Both of the two above-mentioned methods have been heavily criticised, the basket
method for being normative and reliant to some extent on what its critics call 'expert
imperialism', and the proportion method for its failures simply to grasp and explain a large
chunk of any given reality. For example, in Great Britain Margaret Thatcher argued, fiercely,
that according to the statistics, there were inequalities in Britain, but, at the same time, there
was no poverty (J. Klugman, J. Micklewright, G. Redmond 2002). In the case of exclusive
deployment of relative measures a situation may appear in which the proportional growth
(or fall) in average income may not lead to drop in the poverty rate. As Walter Kraemer
wrote in one of manuals, relative poverty is like a ship in a sluice: "... despite raising of the
water level, the submerged part of the ship remains the same." (Kraemer 2000, p. 30).
Indicators of relative poverty are applied in cross country comparison studies (e.g. in the
framework of Eurostat), bearing in mind that the levels of poverty in the compared countries
vary and that the poverty rates are measures of inequality rather than measures of poverty.
In cross country comparisons measures of absolute poverty are also applied. These are
based on some permanent levels of consumption or income which express a concept of
indispensable consumption, e.g. in World Bank research the indicator of 2 (or 4 in 1996 PPP)
dollars per person daily is used as the absolute poverty line.
Subjective poverty
Analyses of poverty have moved increasingly towards accepting some of the definitional
properties and arguments developed by subjective measures of poverty. The premise of this
group of definitions is a liberal treatment of individual feelings and opinions as having a key
significance in defining needs.
In social policy practise particular country specific approaches are applied which respect
both the tradition of national social institutions and eligibility for benefits, as well as the
analytical achievements of local experts in poverty research.
1.2.2. Measures of Poverty Applied in Poland
In analyses of poverty in Poland, both official and those of various experts, many
measures have been applied. This stems from the relatively rich research tradition in this
field.
With tradition in mind, one should begin by looking at the criterion for determining the
so-called 'social minimum'. In Soviet block countries a special category of 'minimum material
security' was used (minimum meterialnoj obespieczennosti). This category didn't exactly define
poverty but rather the threshold of poverty. Those whose income fell below this level were
defined as maloobespieczennyje, or with "little security" (McAuley 1979).
23

Minimum material security in Poland was defined by Tymowski (1973) and Deniszczuk
(1978) as the 'social minimum'. In 1981, under the pressure of just created independent
trade union (Solidarnooeæ) this category was recognised by the government as an official
measure for monitoring living conditions. The social minimum describes the indispensable
level of consumption determining social participation and social integration, which
demands satisfying not only basic needs but also certain other needs beyond them. This is
reflected in the contents of the basket of goods and services indicated as 'basic' for
satisfying needs on a social minimum level. The contents of the basket allow participation
in social life: work, children's education, family life and socialising, participation in culture
– all these on a modest level.
The basket of goods and services was recognised by specialists (doctors, diet experts,
social workers, statisticians, consumption researchers) as indispensable for the normal
functioning of a human being in society.
Since 1982 the social minimum has been estimated by IPiSS and since the beginning of
the 1990s published in the quarterly journal Polityka Spo³eczna [Social Policy].
The social minimum is de facto not a category of poverty, with a level fluctuating around
the average expenditure level in households. Trade unions, however, have used the well-known
term as a synonym for poverty in various populist statements.
The social minimum was never used as a criterion for categories applied in social policy,
such as the minimum wage, minimum pension or threshold income in the social assistance
scheme. However, banks have been known to use information from the social minimum to
assess households' creditworthiness and courts have asked for social minimum information
in order to adjudicate in alimony cases.
A new category of poverty, the so-called subsistence minimum, based on the budget
standard approach, was defined in 1995. This category is used to define absolute, deep-seated
poverty. The subsistence minimum was estimated by experts from the IPiSS on the
basis of the very low family budget costs. This level of expenditure accounts exclusively for
those needs the fulfilment of which cannot be postponed. Any lower level of consumption
leads to physical collapse/illness, etc. The subsistence minimum level is more than twice
lower than the social minimum.
The relative poverty line is equal to 50% of average expenditure calculated for one
consumption unit, according to the OECD's equivalence scale3 (the unit weights are as
follows: 1.0 for the first person; 0.7 for the second and next over 15 years old; 0.5 for each
child). The premise of this definition is, on the one hand, the acceptance of the subjectivity
aspect of poverty (dependency on the average living condition in the country), on the other
hand, pointing to inequalities as the main indicator of poverty.
24
3 Eurostat is currently proposing the adoption of the line at the level of 60% of the median of national income with the
adoption of new equivalence scales.

A measure of poverty based on subjective evaluations has also been defined and estimated. In
Poland the so-called Leyden (Subjective Leyden Poverty Line – SLPL) method has been defined,
systematically estimated and applied by GUS since the beginning of the 1990s. It was determined
on the basis of the household budget survey question: "what is the minimum amount of income
that your family, in your circumstances, needs to be able to make ends meet?" The responses are
measured as a proportion to the actual income in the household. A regression line plotted against
the responses and taking into account family characteristics gives a level of income which coincides
with actual income. That is the basis for the evaluation of "merely sufficient" income (Podgórski
1994). "Taking into consideration the current condition of the household, evaluate the level of
income which would enable living on an average, satisfying level, and what income would not
enable maintaining the household even on the lowest level" (Podgórski 1991).
To work out the official poverty line in Poland in the first half of the 1990s one needs to look
at the income threshold entitling social assistance allowances. This income threshold was
defined by the category of the minimum pension, determined as 35% of the average wage.
Income per person entitled to social assistance could not exceed 90% of the minimum
pension. In the regulations the minimum pension was never officially classified as the poverty
line, but social policy from that time indicates that the minimum pension was used as a
screening device to separate applicants who needed support from those who did not. It means
that in practice the minimum pension determined possibilities of obtaining social assistance and
housing allowances (Topiñska 1997). The World Bank's poverty studies on Poland were made
with the minimum pension serving as the official poverty line (World Bank 1994 and 1995).
Since 1996 the income threshold has been explicitly defined in the law on social
assistance. Henceforth the income threshold can be recognised as the official poverty line.
The level of this income threshold was set as 35% of the net minimum wage in the starting
point (1996) and was indexed against price increases; once or twice a year depending on the
rate of price growth. Currently the income threshold is equal to approx. 20% of average
earnings.
According to GUS and Eurostat practice, relative poverty lines are systematically applied
in Poland, as 50% of the average expenditure basket estimated on the basis of HBS.
However, there are several doubts associated with applying this poverty line. The main
question is if the relative income approach is really a poverty measure? Some argue that
relative poverty is a misleading and muddled concept (Lipton 1995). The main disadvantage
of a pure relative measure of poverty lies in the fact that even in a situation of proportional
nominal increase in expenditure (or income) the level of poverty remains unchanged.
In the World Bank's studies the so-called two-(or four-)dollar-a-day poverty line is used
(actually $ 2.15 per person per day or $ 4.30 in the 1996 PPP).
The relatively wide scope of poverty studies in Poland has meant the use of many poverty
definitions. The definitions, poverty lines and institutions providing them are showed in Table 1.6.
25

Table 1.6. Poverty lines applied in Poland
1.2.3. Research into Poverty
Poverty has for some time been always a closely studied area of research in the social
sciences in Poland. Various studies were conducted during the communist period, though
their findings were either not published at all or were only partially published. The first
studies appeared in the 1970s (a period of considerable political liberalisation in Poland) and
developed during the 1980s. They were important for developing the methodology of
poverty research. These studies provided insights into environments that were at that time
euphemistically referred to as 'spheres of deficiency' (Jarosz, Fr¹ckiewicz 1984).
World Bank experts have also taken an interest in the problem of poverty in Poland. They
conducted initial research into the prevalence of poverty in the last few years of the 1980s
(World Bank 1990).
A relatively rich vein of poverty studies developed in the 1990s, with various types: some
aiming to improve methods for poverty research, some to collect better data to be used to
identify poverty, others to develop theories explaining the causes and cases of poverty
(sociological research), and others that aimed explicitly to develop methods for fighting poverty.
Types of research and authors in the 1990s
Research to date can be generally divided into qualitative and quantitative camps. The
former have been to a larger extent conducted in the framework of sociological research.
The latter are most frequently used by economists and statisticians, although interesting
quantitative analyses have also been made by sociologists.
26
Categories of poverty
lines
Lines applied Institutions and authors using
given poverty measurement
Absolute poverty Subsistence minimum, Two - dollar-a-day
Four - dollar-a-day
IPiSS, GUS, the World Bank
Relative poverty 50% of the average household
expenditures per consumption unit
(OECD equivalence scale)
GUS
Subjective poverty Leyden Poverty Line GUS (Podgórski)
Income threshold in
the social assistance
institution
447 PLN with application of OECD
equivalence scales
MPiPS; social assistance
centres
Threatened with
poverty
Social minimum; basket of goods and
services essential for participation in social
life (assuring integration)
IPiSS, GUS, trade unions
Source: Own comparison.

27
Author/institution
undertaking research
Subject and Scope of Research Data Source Period
Qualitative Researches
Tarkowska, E., Zrozumieæ
biednego (To Understand
the Poor) IFiS PAN
employees of the former state
farms (PGR)
interviews in the
field
1997
Tarkowska, E.
/Korzeniewska (ISP)
children of the former state
farms’ employees
interviews in the
field
2000
Tarkowska, E./Laskowska -
Otwinowska, J.,
Fodor, E./Domañski, H.
Poverty Ethnicity & Gender
IFiS PAN
Romans in Spisz
Feminisation of poverty in the
countries of Central and Eastern
Europe
interviews
analysis of a
representative
sample
2001
2000
Warzywoda-Kruszyñska,
W. with a team, ¯yæ i
pracowaæ w enklawach biedy,
(¯yæ) na marginesie
wielkiego miasta (To Live
and Work in Enclaves of
Poverty, (To Live) on the
Margins of the Big City)
Uniwersytet £ódzki (£ódŸ
University)
enclaves of urban poverty in
£ódŸ
group interviews
with employees of
social institutions
1993
1994
1997-1999
Zab³ocki, G. with a team,
Ubóstwo na terenach
wiejskich Pó³nocnej Polski
(Poverty in Rural Areas of
Northern Poland)
Uniwersytet Toruñski i
Bank OEwiatowy (Toruñ
University and the World
Bank)
“poor” gminas in Koszaliñskie,
S³upskie, Pilskie, Elbl¹skie,
Olsztyñskie and Suwalskie
voivodships
interviews with
residents, analysis
of official
documents and
interviews with
key persons
1997
Rossa, J.
IPiSS and IPS UW
workers’ quarters of Silwan and
Stilon plants in Gorzów
Wielkopolski
interviews with
residents of
workers’ quarters
1998
Frieske, K. W./Polawski, P.
/Zalewski, D.
small towns – seats of gminas expanded
interviews with
selected families
supported by
social assistance
1997
1998
Quantitative Researches
The World Bank
Milanovic, B.
Topiñska, I.
estimation of poverty rate and
poverty gap
HBS 1990
1994
1995
The World Bank
Okrasa, W.
dynamics of poverty and
research on effectiveness of the
social safety net
panel 1993-1996 1993-1996
GUS
Szukie³ojæ-Bieñkuñska, A.
Podgórski, J.
estimating rate of relative
poverty and lines established by
IPiSS (subsistence minimum and
social minimum),
estimating subjective poverty
HBS systematically
since 1993
systematically
since 1990
Table 1.7. Comparison – research into poverty in Poland during the transition period

28
Table 1.7. Comparison – research into poverty in Poland during the transition period – cont.
Author/institution
undertaking research
Subject and Scope of
Research
Data Source Period
Panek, T./Kotowska, I.
SGH – Warsaw School of
Economics
evaluating the risk of
deficiency
special random sample
within the framework of
the project „Diagnoza
2000” /Diagnosis 2000/
Feb – March
2000
CASE / ZEW
Beblo M.
Golinowska, S/Laner C./
Piêtka, K. /Sowa, A.
intergenerational poverty
dynamics
HBS, LFS and POMOST 2000
IPiSS
(Sajkiewicz B./Kurowski P.)
valorising the value of
subsistence minimum and
social minimum
records of prices
monitored by GUS
since 1981 in
the case of
social
minimum and
since 1996 in
the case of
subsistence
minimum
Quantitative – Qualitative Researches
IPiSS – Polish Poverty I
and II
Golinowska, S./Kordos, J.
complex analysis of
poverty
HBS 1996
1998
IFiS PAN
Beskid, L.
common perception of
poverty
surveys by IFiS PAN and
CBOS
1989 – 1995
Research of
Methodological
Character
Panek, T.
SGH
application of
multidimensional analysis,
monitoring the poverty
sphere by the method of
multidimensional analysis
HBS
HBS
1993-1994
May 1995-
November
1996
Szulc, A.
SGH/GUS
equivalence scales HBS 1993-1994
1995-1996
IPiSS
PBZ Polish Poverty I
(Deniszczuk, L./
Sajkiewicz, B.; co-ordination
Golinowska, S.)
defining and estimating the
category of subsistence
minimum by the basket
method
verification of the social
minimum basket
contents of the basket
described by experts
1995
1995
Source: Own evidence.

Research problems
Quantitative research requires representative data bases. Results of research on
household budgets (conducted in Poland systematically since 19744) can be treated as
such. This research enables cross-sectional and over-time analyses. These data bases are
not yet representative enough for margin income groups or for affluent ones or for the
poor. Moreover, they are not suitable for the application of dynamic approaches. Dynamic
analyses require panel research, that is the participation of the same respondents over
several years. It is not easy when research samples are created by rotation method, which
is most frequently used for household budgets' data bases. The first panel enabled a two-year-
long analyses (in the years 1990/1991, 1991/1992). The longest panel research
covered a four year period (1993-1996). Talking about the significance of data sources
needed for dynamic approaches, one should note that a considerable part of the
coefficients proposed within the concept of open co-ordination of social policy within the
EU is based on panel analyses (Atkinson 2001).
Only the dynamic approach allows for verification of the concept of poverty as a
temporary phenomenon. This thesis acts as a premise for further research (Benjamin
Rowntree, is a kind of founding father of this approach), and even more importantly, for
many policies aimed at counteracting and combating poverty. The thesis proposes that the
risk of poverty is connected with particular stages in the cycle of human life. Two stages are
the most susceptible to poverty: the early period of family life – the period when children
are born and taken care of, and later life – the period of old age.
Another concept – in direct contradiction to the last one – holds that the phenomenon
of poverty is of a constant character and concerns certain groups which pass the culture
of poverty down from one generation to the next. According to some sociologists, these
groups form a class, or more precisely an underclass, which in many countries has an
ethnic aspect. A classic description of the underclass is the work of Oscar Lewis on the
culture of poverty of Mexican and Puerto Rican immigrants in the USA (Lewis 1959 and
1966). The culture of poverty in Poland does not have an ethnic aspect, but there are a
few marginalised categories described by sociologists as "candidates for the Polish
underclass" (Tarkowska 1998).
Verification of each of the concepts requires different types of research. In the case of the
first, quantitative dynamic approaches are necessary; in the case of the second qualitative
research is required.
29
4 Hitherto they have only been conducted for the employed in the state sector and then not systematically.

1.3. The Picture of Poverty in the 1990s
Looking at the research findings on poverty today, it is natural to wonder about the period
proceeding the transition decade. Did poverty exist in the communism period, or not? Studies
from the 1970s and 1980s indicate that poverty did exist at that time, though was frequently
related to family dysfunction, substance abuse, serious illness and old age. Poverty did not
appear among employed persons or their families since full employment guaranteed
subsistence, however modest and uniform. According to GUS estimations related to so-called
'low income' (similar to the social minimum category), one can assume that Poland began the
transformation with approximately 20% of its population threatened with poverty (Kordos
1992), though not a deep one.
Table 1.8. Poverty rates (poverty headcount) according to different poverty lines – people living at the
given poverty level and below this level, as a % of the total population
In the 1990s we can see two periods with the significant increases in poverty. The first
period, in which poverty rose suddenly, was immediately after the collapse of the communist
block and the beginning of the transition from the planned to the market economy. In the
space of only two years, 1990-1991, GDP fell by approximately 18%5.
The second period is the end of the 1990s. Since 1998 economic growth has slowed and
unemployment grown continuously.
Of the countries in the region Poland comes in fourth behind the Czech Republic,
Slovakia and Hungary in terms of the scale of poverty. Polish poverty rates are similar to
those in Bulgaria. Other countries in the region have much higher indicators.
30
Poverty lines 1992 1994 1996 1999 2000
World Bank two-dollars-
a-day
1.2 (1998)
Subsistence minimum - 6.4 4.3 6.9 8.1
Relative poverty line* 12.0 (1993) 13.5 14.0 16.5 17.1
Income threshold in the
- - 13.3
social assistance scheme
(1997)
14.4 15.0
World Bank four-dollars-
a-day
10.0 (Milanovic
study 1993-1995)
18.4 (1998)
Subjective** 32.6 (1990) 33.0 30.5 34.8
Social minimum 32.4 47.9 46.7 52.2 54.0
Notes: * 50% of average expenditure per OECD consumption unit,
** % of households at the level and under the level of poverty.
Source: GUS, IPiSS, World Bank (2000).
5 Estimation concerning the drop in GDP differ, even in the yearbook of GUS for subsequent years; generally, the later the
yearbook, the smaller the decrease. The figure of 18% is the most widely used.

31
Chart 1. Poland: Poverty rates in 1994 – 2001, according to different measures
8.1%
33 .0% 30.8%
6.4%
9.5%
4.3%
5.4% 5.4%
5.6% 6.9%
17.0%
12.8% 14.0% 15.3%
15.8% 16.5%
13 .5%
17.1%
13.6% 15.0%
13.3% 14.4%
12.1%
30.8% 30. 5% 30.8%
34.8% 34.4%
32.4%
47.9% 46.7 % 46.7%
50.4% 50.0%
52.2%
54.0%
57.0%
60%
40%
20%
0%
1994 1995 1996 1997 1998 1999 2000 2001
Subsistence minimum Relative poverty line Social assistance treshhold
Subjective line Social minimum
Table 1.9. Rates of poverty in Central and Eastern European countries
Country Relative Poverty* Absolute Poverty**
Czech Republic 1.3 0.8
Slovakia 1.7 8.6
Poland 12.0 (1994)
15.8 (1998)
18.4
Hungary 9,9 15,4
Bulgaria 15,0 18,2
Rumania 39,0 44,5
Estonia 37.0 19.3
Latvia 22.0 34.8
Lithuania 30.0 22.5
Belarus 22.0 10.4
Moldavia 66.0 84.6
Russia 22.8 50.3
Ukraine 63,0 29,4
Notes: * Adopted poverty line – 50% of household's expenditure calculated per one consumption unit according to OECD
equivalence scales, data from years: 1993-1996;
** Poverty line used by World Bank – 4,30 USD per person daily, data from years 1997-1999.
Source: ICDC Uniceff 1999, Flemming and Micklewright 1999, World Bank 2001, the Author's complements.

1.3.1. Most Vulnerable Groups
Statistical evidence helps in pinpointing who the poor are. Using socio-economic criteria
to classify household groups in the sample of the household budgets survey we can see that
families living from non-earned sources (unemployment and social assistance benefits) are
the most poverty-vulnerable household group. In 2000, in 7% of the total household sample
all household members were jobless.
Unemployment as the main cause of poverty
At the beginning of the transformation, 75% of the poor – according to the official poverty
line (minimum pension) – were stricken with unemployment (World Bank 1994). Very soon
after the emergence of open unemployment in Poland, certain social groups appeared to be
disproportionately affected. They included young people, unskilled workers and residents of
small towns and villages. At the end of the 1990s unemployment was still a key cause of
poverty, but was no longer on such a high scale. The most affected group by unemployment
are again young people (below 25 years of age). In general, young people is the segment most
exposed to the risk of unemployment. This is a very common feature for many countries
suffering from severe unemployment. In Poland, however, the problem of youth
unemployment is now much more intensive because labour supply inflows are very high.
Women were the second group most strongly affected by unemployment, in spite of
their higher levels of education. The position of women depends very much on the structure
of employment by sector and the contribution of male- or female-dominated industries to
employment. So-called "female" sectors in Poland (light industry, pharmaceutical industry,
and education) are just beginning to see deep cuts and restructuring.
32
Table 1.10. Unemployment rates according to differently affected groups by unemployment, in %
Most affected
groups
1993
November
1996
November
1997
August
1999
February
2000
1st quarter
2001
1st quarter
Average 14.9 11.5 10.7 12.5 16.7 18.2
Young people
31.6 26.2 23.5 28.5 37.9 41.2
(15-24)
Unskilled (basic
and lower
education) workers
15.0 12.9 12.5 11.7 22.1 22.9
Women 16.5 13.4 13.1 13.5 18.5 19.8
Urban population 16.9 12.0 11.5 12.9 17.1 19.2
Elderly (above 45) 8.5 6.5 6.3 6.3 10.6 11.9
Source: GUS: systematic LFS.

Low employability
According to qualitative research on poverty in Poland, the main features of the exclusion
process are: long-term unemployment and passivity towards job seeking and/or inability to
work in sectors of the economy where demand is highest. There are several overlapping
reasons for low employability: low human capital, living in non-urbanised areas, alcohol abuse
or chronic illness and disability. If we look, for instance, only at social assistance recipient
groups with an alcohol dependency problem, we can see additional exclusion-bred
problems: loneliness and rejection by families due to time in prison or a long-term absence
from the family, for example due to job migration (in-country or abroad). The overwhelming
majority of those people are men.
Table 1.11. Long-term unemployment: the share of the unemployed over one year in total
unemployment, %
Rural poverty
The worsening of the situation in agriculture is the second major cause of poverty in
Poland. Income from agriculture largely collapsed due to the withdrawal of input subsidies
and more liberal import policies. The explicit government policy of ensuring rural-urban
income parity was ended. At the end of the 1990s disposable real income in the farmers'
households declined substantially (by about 30% in 1998-2000), whereas in other household
groups the figure increased several percent.
As a consequence, people who live in the countryside and have some link with agriculture
(belonging to farmer- or mixed worker-farmer-households) are poor more frequently than
the average. Almost 40% of peasants in small farms (under 10 ha) are poor (GUS 2000). But
one should also note that the range of income among farmers in Poland is larger than among
non-farming employees (Deniszczuk 1998).
Young families poorer
Studies on poverty in Poland carried out in the 1990s showed fundamental changes in
income redistribution policy across generations. These changes were advantageous for older
generations and disadvantageous for younger generations and children. The syntheses of
33
Most affected groups
1994
November
1997
August
2000
(st quarter)
2001
(st quarter)
Average 22.7 39.2 35.1 40.2
Elderly (more than 45) 41.4 47.9
Women 43.7 38.0 44.0
Source: LFS.

results using various poverty lines and definitions (absolute, relative, and subjective)
confirmed that households with the highest poverty risk are large families (Golinowska
1996). With each line applied, the poverty rate for families with three or more children is the
highest, much higher than for an average family, single-parent families, or pensioners. Poverty
rates for families with four or more children are dramatic. In addition to their worse income
situation, the health of 20% of persons in these families is endangered, and as much as 85%
are below the social minimum.
Table 1.12. Poverty rate among households by different socio-economic groups in 1999-2000
1.4. Concluding Remarks
Poverty in Poland during the 1990s more closely resembled poverty in European
countries than in other regions of the world, such as Asia or South America. Its main trait is
the higher probability of falling into poverty among young people and families with numerous
children. The older generation, taking advantage of an advanced social insurance system and
democratic political structure, can in the main effectively protect itself from poverty.
34
Type of households
Subsistence
minimum
Relative
poverty lines
Subjective
poverty lines
Income
threshold in
the social
assistance
Social
minimum
Year 1999 2000 1999 2000 1999 2000 1999 2000 1999 2000
Total 6.9 8.1 16.5 17.1 34.8 34.4 14.4 13.6 52.2 53.8
Employees 4.7 5.8 12.4 13.6 25.3 22.6 10.7 10.4 51.2 52.5
Farmers 13.3 12.9 29.9 26.2 44.0 43.6 26.8 21.3 68.9 68.5
Employees –
9.2 8.6 24.7 20.9 28.9 25.6 21.3 16.2 64.8 62.9
Farmers
Self-employed 3.2 3.7 8.2 8.5 16.4 17.1 6.4 6.5 36.6 38.2
Old age and disability
6.1
8.0
14.3
16.6
-
-
12.5
13.3
44.0
pensioners
- old-age
ensioners
3.6
5.1
9.6
10.8
34.9
38.2
8.6
8.4
35.2
48.3
39.0
Maintained from
other non-earned
sources
28.4 30.4 46.9 45.5 78.6 72.5 43.8 30.4 78.4 80.3
Source: GUS 2000.

Polish poverty remains related to the structural backwardness of the economy –
relatively high in agriculture, and deterioration in industry, as well as to the high regional
disparities. The restructuring process of the economy intensified in the second half of the
1990s and brought many difficulties in living conditions, especially for those groups most
strongly affected by the changes. At the same time, a large number of threats potentially
affecting young people emerged on the labour market. This high labour supply at the end of
the 1990s and beginning of the 2000s is not being met sufficient by available jobs. This in turn
makes young people susceptible to poverty. However, the rate of poverty among children
depends to a large extent on social policy measures. Social policy in Poland today does not
give priority to children's well-being. Perhaps poverty research findings will contribute to
changes in social policy goals.
Poverty in Poland is characterised by four different main features. One is connected with
the lack of jobs and is the most important challenge facing economic and social policy. The
second results from the lack of effectively operating social policy towards families with
children. This policy defect concerns mainly benefits for families and the social functions of
the educational system. The third is linked to agriculture and life in rural areas. On the one
hand this is a problem of on-going economic neglect, on the other hand, little possibilities and
low adjustment abilities of the rural population to dealing with competition in the open
market economy. The fourth is formed by marginalised individuals and groups. This requires
plenty of work to bring them back into work and social life.
A strategy that would be able to deal with such a diverse range of poverty elements itself
requires a diversified range of activities across different fields. Such a strategy should not be limited
to a narrow social program. Also demands a wide-ranging economic anti-poverty programme.
35
Table 1.13. Poverty among vulnerable groups of population in 2000, % of persons below the poverty line
Specification Relative poverty line
Official poverty line –
social assistance
threshold
Average for the population 17.1 13.6
Living from unemployment and social
49.1 43.1
assistance benefits
Households with at least one unemployed
person
36.0 -
Households whose head has low education
– primary and incomplete primary
28.9 23.9
Rural population 25.8 20.6
Couple with
3 children
24.5
19.3
4 and more children
47.9
40.8
One parent family 19.2 15.3
Source: GUS 2001 according to HBS.

Introduction
The effective allocation of social transfers is one of the main dilemmas facing social policy.
In Poland, where cash transfers as a share in GDP are particularly high, concern in this area
is justifiably high. The objective of social transfers is to replace or supplement earned income.
Therefore, modern social policy has to be considered in connection with the labour market
policy. One of the indications of the effectiveness of social transfers, other than retirement
pensions, is their labour supply effect: Are social transfers well targeted or do they provide
excessive financial comfort and therefore create disincentives to acquire work? The
consequence of this trap can be long-term unemployment or/and falling into poverty if the
level of transfers is not high enough. On the other hand, from the public expenditure
viewpoint it creates a long-term financial burden.
This paper is an empirical study on the labour supply effects of social transfers in Poland.
Most international empirical works related to the labour supply focus on its links with the tax
system. The most prominent programmes aimed at reforming the welfare system in the US
(Earned Income Tax Credit) and UK (Working Families Tax Credit) were related to these studies
but also inspired new research. However, the analyses of the relationship between social
transfers and the employability of recipients has been, to a certain extent, under-researched.
In Poland, social transfers and taxes were usually analysed as a subject of redistribution, and
not in the context of the labour market. Several studies (Topiñska 1991, Nage, Neneman
1995, Kramer, Neneman 1995, Górecki, Wioeniewski 1998, Topiñska, Styczeñ 1999, Aksman
2001) were inspired by questions about the redistributive character of transfers and taxes.
They generally have shown that there is a classical model: taxes are paid by more affluent
37
Part 2
Labour Supply Effects of Social
Security Transfers
Katarzyna Piêtka*
* I would like to express my thanks to Prof. Stanis³awa Golinowska for her professional assistance, and to Miriam Beblo
and Charlotte Lauer for their technical advice.

households, while transfers assigned to poorer families. The end of the 1990s demonstrated
higher imbalances in public finances and on the labour market. This led to a new set of
questions about the role of transfers and taxes in the context of labour supply and demand.
This study focuses on one of these issues – labour supply effects of social transfers. But it
also addresses the question concerning the methods of addressing poverty. Are transfers an
effective measure to help the poor?
The study is based on the Labour Force Survey (LFS) individual data. The methodology
applied in the research was obliged to use a static econometric model since the scope of the
LFS and the time-frame of the panel data were insufficient for a dynamic approach. The
dynamic aspect of the research was introduced by applying the model at two different points
in time: the 4th quarter of 1997 and 2000. Although the static approach does not enable us
to draw concrete conclusions as to the causality at play in the relationship between transfers
and labour supply it does show a clear relationship between them. Seeking work is positively
correlated with receiving transfers. This outcome of the econometric model had already
taken into account all the other (11) available individual characteristics that could have had a
separate effect on the labour supply. Moreover, an additional presented analysis indicates
that the level of transfers is relatively low and may in fact push those transferees who do not
supplement benefits with earned income into poverty.
Since the outcome of the research cannot provide clear answers to questions related to
the effectiveness of social transfers in Poland, any policy recommendations that stem from
them have had to be equally circumspect. Various policy ideas are presented in the
Concluding Remarks.
2.1. Background
2.1.1. Review of Social Policy in Poland During the Transition
The implementation of the Balcerowicz Plan in 1990 put Poland on a two-pronged
transition course: introducing a market economy, while at the same time tackling problems
thrown up specifically by the transition itself. As in many other socio-economic spheres,
social policy had to address the problems found in all market economies, such as
unemployment, poverty and income disparity, replacement of the traditional family model of
male 'bread winner' with a larger role played by women in the labour market, and the co-existence
of public and private expenditures to achieve social goals. Additionally, the
transition process from centrally planned to market economy, which implied social changes,
provided separate challenges for policy.
38

As Golinowska (1996, 2000) has argued, social policy over the last decade has passed
through various stages. In the years 1990-1991 the central effort went into replacing price
and wages controls, and replacing the existing policy of full employment with a rational social
policy devoid of economic parameters. This required liberalisation of prices and wages1, and
the building of institutions and mechanisms in the labour market and in the area of social
assistance. A substantial shift from an economic-based social policy to an income-redistribution
policy produced some unwanted serious results, such as the absorption of
many unemployed into the old-age and disability pension system.
The next two years, 1992-1993, were devoted to correcting earlier over-generous social
expenditure decisions. These were mainly in the area of lowering the average level of social
transfers, including unemployment and family benefits, as well as price indexation of
pensions. Additionally, the eligibility criteria for some of the benefits were sharpened.
The years 1994-1995 brought another shift in social policy. High economic growth and
improvement in the public finances were conducive to higher social expenditures. Separate
programs were established to develop social assistance. However, besides rises in the level
of benefits and pensions and new targeted measures (including family benefits), various
programs aimed at increasing activity of beneficiaries on the labour market were undertaken.
The period 1996-1998 was devoted to preparation of systemic changes. A partial reform of
the disability system in 1997 redefined 'inability' in 'a biological sense' as 'the remaining capacity to
work' and strengthened eligibility controls. Besides this, certain restrictions on the transfers/wage
ratio were imposed by replacing the wage indexation of social transfers with mixed (price and
wage) indexation. This was an important change also from the fiscal policy viewpoint.
In 1999, four big reforms in the social policy sphere were launched. (1) Old-age pension
reform introduced a new obligatory pillar with pension funds in addition to the diminished
pay-as-you-go system. In parallel, pensions within PAYG were linked tightly via the formula
with contributions accumulated throughout a whole working carrier (Notional Defined
Capital). The NDC mechanism is supposed to make the system self-sufficient and stable. This
would mean that the pension system is devoid of any distribution mechanisms and practically
excluded from the area of public policy and politics. (2) Health reform aimed at introducing
so called regulated market with self-governance health institutions under public financing.
Health expenditures were always very low in Poland. The reform was expected to push for
efficiency of health services providers and better access to health care. The first experiences
are showing however, that the institutions hardly increased their efficiency and, instead, high
transaction costs and growing liberalisation on that market limited resources for the pure
health services. There are signs that it eventually limits the access of population to basic health
care and increases final costs of the treatment or creates additional demand for resources in
39
2. Labour Supply Effects of Social Security Transfers
1 A tax on excessive wage increases (so called "popiwek") was imposed on the wage bill of companies (and not individual
remuneration) and was introduced only temporarily as an anti-inflationary measure.

other spheres of social policy. (3) Education reform aimed at adjusting the education system
to the requirements of the modern labour market. (4) Decentralisation, including shifting
some social policy areas to local administrations in order to increase their efficiency.
Pension reform assumed also a gradual cancellation of early retirement. It should bear
fruit in several years, although its immediate outcome was to increase the number of pre-retirement
pensions (introduced in 1997). This is a way of protecting older workers against
unemployment at the age just prior to retirement.
Family policy during transition period has rarely enjoyed a clear consensus as to aims or
actions. Ultimately, protection of the family against poverty was and remains its only plank,
but this is very vague and negligible (Golinowska 2000, p.112).
2.1.2. Social Expenditures
Social expenditures are part of budget expenditures, though are not clearly defined either
in the budget law or in the statistics. The United Nations Organisation proposes criteria
which allow a separation of social expenditures from overall budget spending:
– expenditures with the purpose of income distribution,
– public expenditures targeted at households,
– public expenditures earmarked for closely defined social goals, regardless of the source
and form of financing (Golinowska 1998).
Following these three criteria, Golinowska and Hagemejer estimated the level of social
expenditures in Poland in 1990-2001 (Table 2.1).
Total social expenditures over the period of transition grew to a high level of over 30% of
GDP, and only in 2000 lowered to 27%. Their dynamics reflect changes in social policies. In
particular, the rapid growth in cash transfers during 1990-1992 illustrates the favourable
indexation (despite high inflation) and liberal access to transfers during the first years of the
transition. Their share in GDP doubled during the first three years of the transition (from 11%
in 1990 to 20% in 1992). Some corrections in the system over the subsequent few years
caused a slowdown in their real growth to 0.5-2.5% in 1993-1994. Starting in 1995, in parallel
with high economic growth and a return to generous social policies, cash social expenditures
kept growing in real terms until 1997. The bulk of this growth resulted from old-age and
disability pensions and protection for elderly workers against unemployment. Since then the
transfers have tended do diminish in general at the expense of payments other than pensions
and unemployment benefits. 1998 brought significantly slower increases in cash transfers due
to rigorous increases in all kinds of transfers and, additionally, lowering unemployment. The
transfers decreased in real terms over the next 2 years due to savings on social assistance and
40

Table 2.1. Social transfers, including cash transfers to households
other forms of cash support (e.g. subsidies for house utilities). The real increase in cash
transfers in 2001 resulted from the delayed indexation of pensions that had been due in 2000.
In 2001, they reached the level of 18.2% of GDP.
Poland's high level of social expenditures puts it among countries with the highest income
redistribution (Golinowska 1998). However, these countries (like Denmark, the
Netherlands, Belgium, Sweden and Germany) are at present at a much higher stage of
economic development than Poland.
Cash transfers include all transfers directed to households (old-age, disability and family
pensions, family benefits, unemployment benefits, social assistance benefits and scholarships,
as well as some other discretionary payments). Old-age, disability and family pensions have
the largest share in cash transfers – 76% in 1998 (Table 2.2). Second place belongs to family
benefits, with a 5.8% share. Other transfers are very small, with social assistance at a very
low level of just above 2% of all cash expenditures. This structure shows that high social
expenditures in Poland in the part of cash transfer are composed mainly of pensions, and not
social transfers based on poverty or other criteria of social non-insurance policy.
2.1.3. Individual Transfers to Households
The main kinds of social transfers to households include:
Old age pensions – currently paid in the PAYG system. The pension is provided to all
insured persons who fulfil age, or age and working record, requirements. The legal
retirement age is 60/65, although until 1999 the early retirement age was 55/60. In the
employee insurance system (ZUS), disability pensions have been automatically changed into
retirement pensions at the legal retirement age. In the farming system (KRUS) disability
41
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
social
expenditures
bn z³oty 14.6 24.6 37.6 48.6 70.2 97.9 129.2 156.0 171.7 n.a. 184.8 n.a.
% GDP 24.6 29.9 32.9 32.6 34.1 31.8 33.3 33.0 31.0 n.a. 27.0 n.a.
real growth, % - -0.8 6.9 -4.6 9.3 9.2 10.0 5.1 -1.6 n.a. n.a. n.a.
cash transfers
bn z³oty 6.4 14.0 22.8 31.0 42.0 61.2 76.3 90.6 101.8 108.8 116.9 131.4
% GDP 10.8 17.0 20.0 20.8 20.4 19.9 19.7 19.2 18.4 17.7 17.1 18.2
% of social
43.6 56.9 61.5 63.8 59.8 62.5 59.1 58.1 59.3 n.a. 63.3 n.a.
expenditures
real growth, % - 28.4 13.9 0.5 2.5 14.0 4.0 3.3 0.5 -0.4 -2.4 6.5
Source: S. Golinowska, K. Hagemejer 1999, Golinowska 2002, own compilation based on MPiPS 2002.

Table 2.2. Structure of cash transfers in 1990, 1995, 2000, and 2001, %
pensions are not transferred into old-age pensions. In 2000, the average level of an employee
retirement pension was 60% of the average wage in the economy (ZUS 2001).
Disability pensions – until 1997 assigned to people based on their state of health; since
late 1997 – based on their remaining capacity to work. There are 3 groups of invalidity: III-group
disability represents partial impairment (or since 1997 – partial inability to work); II-group
disability – complete impairment (since 1997 – inability to work); I-group disability –
complete physical (work) inability and independent existence incapacity. I- and II-group
disability pensions are assigned permanently. The III-group pension can be withdrawn if,
based on the periodical medical tests, an improvement in physical (work) capacity can be
assessed. The average duration of the III-group entitlement is around close to 5 years (ZUS
2001); for all the groups together – 13.5 years. In 2000, 50% of disability pensions in the
employee system were III-group pensions, 37% – II-group pensions, and 13% – I-group
pensions. In 2000, the average level of a disability pension was 44% of the average wage in
the economy (ZUS 2001).
Family pensions – pensions from the employee insurance system inherited by family
members of the insured family provider (to orphaned children or widowed spouses). The
period of entitlement for children is until adulthood or the end of their studies; for spouses
it is the moment of retirement. In 2000, the average family pension was 52% of the average
wage in the economy (ZUS 2001).
Unemployment benefits – for those insured in the Labour Fund and having an employment
record. The duration of the entitlement period is regionally differentiated between 6, 12 and 18
months. The basic level of benefit is set at 36% of the average wage, and the effective rate differs
between 80-120% of the basic level depending on the working record.
42
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
Cash
transfers,
including
100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0
old-age,
disability and
family
pensions
76.5 75.0 75.0 77.2 78.8 80.5 81.2 75.8 76.3 79.5 79.2 79.3
family
benefits
15.5 14.7 12.5 10.3 8.6 6.2 6.0 5.7 5.8 6.9 7.5 7.3
unemploy-ment
benefits
3.1 7.8 8.3 8.7 8.8 8.8 8.4 5.7 3.0 3.6 4.9 5.5
social
assistance
benefits
1.3 1.4 1.5 2.5 2.0 2.1 2.5 2.3 2.2 2.3 2.2 2.0
scholarships 1.3 0.10 0.06 0.05 0.05 0.05 0.05 0.07 0.09 0.09 0.12 n.a.
Source: S. Golinowska, K. Hagemejer 1999, Golinowska 2002, own compilation based on MPiPS 2002.

Table 2.3. Number of pensioners and unemployment beneficiaries in 1990, 1995 and 2000, in thousands
Other transfers – these include all kinds of benefits and allowances with both entitlement
based on the insurance and not:
– family benefits,
– nursing disabled benefits,
– child-care benefits,
– maternity benefits,
– social assistance benefits,
– sick, rehabilitation, accident, birth, and funeral allowances.
Mixing incomes from different pensions, benefits or work is limited by the income criteria.
Pensions are subject to gradual reduction in case of earnings from work exceeding 70% of the
average wage, until complete suspension when the earnings' level reaches 130%.
Unemployment benefits naturally cannot be combined with earned income. Most of the
benefits are means-tested.
The huge burden of social cash expenditures results mainly from the growing number of
pensioners. During the decade 1990-2000 their number increased by 37%, with retirement
pensioners growing most rapidly (by 44%). The number of unemployment beneficiaries has been
decreasing in recent years since more and more unemployed are losing their entitlements to benefit.
2.1.4. Unemployment in Poland
Since the beginning of the Polish transition employment in public sector enterprises has
been undergoing continuous reductions. By the end of 2000 it had been lower by more than
half compared to the end of 1990. Wowever, this process encompasses also the effects of
privatisation. Growth in employment in the private sector allowed an absorption of a large
portion of this laid off labour force (Table 2.4). Not all dismissed employees became
unemployed. As discussed above, the pension system played an important role in absorbing
specifically older laid-off workers.
43
1990 1995 2000
Retirement pensioners (ZUS+KRUS) * 3 220 4 488 4 630
Disability pensioners (ZUS+KRUS) * 2 628 3 391 3 472
Family pensioners (ZUS+KRUS) * 1 029 1 179 1 310
Unemployment beneficiaries ** 892 1 548 549
Notes: * Annual average;
** At the end of the year, the unemployed with benefit entitlement.
Source: GUS (1997 and 2001).

Table 2.4. Structure of the employment in Poland, end-year, 1990-2000, in millions
Registered unemployment grew, however, by more than the balance of total employment
despite the softening of labour market tensions through the expansion of the pension system
(Table 2.5). This was caused not only by the growing inflow of graduates throughout the 1990s,
but also due to the registration of people who had been outside of the labour force – specifically
at the beginning of the transition. The health reform in 1999 provided additional incentive for
the unemployed to register since the registered unemployed were entitled to health insurance.
Table 2.5. Unemployment in Poland, end-year, 1990-2000, in millions
According to the LFS, the share of the long-term unemployed was over 41% in the 4th
quarter of 2000, although the ratio had been relatively stable since 1992 (when the LFS was
launched). Structural unemployment, which is responsible mainly for long-term unemployment,
started in early stages of the transition.
2.2. Empirical Study
2.2.1. Goals of the Investigation
The empirical study of social transfers' effect on labour supply is based on LFS (4th quarter
1997 and 4th quarter 2000). The aim of the study is to investigate if transfer entitlements
provide disincentives to taking work, in two different points in time. The investigation focused
44
1990 1995 2000 Balance 1990-2000
Public sector 8.6 5.6 4.0 -4.6
Private sector 7.9 9.5 11.2 3.3
TOTAL 16.5 15.1 15.2 -1.3
1990 1995 2000 Balance 1990-2000
Unemployed persons, in millions 1.1 2.6 2.7 1.6
unemployed entitled to unemployment
benefit, % of total
79.2 58.9 20.3 -
Unemployment rate 6.5 14.9 15.1 -
Long-term (over a year) unemployed
according to LFS, % of total 39.9* 39.9 41.1 -
Note: * figure for 1992, when LFS was launched.

on people aged 15 and above (that is those covered by the LFS) in the groups that should be
considered realistic providers of labour and, at the same time, are desired to be active on the
labour market. This meant dropping retirement pensioners and all those above the effective
retirement age2, the I-group disabled3, as well as all kinds of students.
Secandly, there is a group of people who combine earned income with non-earned income
(e.g. wages from part-time jobs with a disability pension), or even two kinds of transfers.
These people were also dropped in order to eliminate cases which do not fall univocally into
one of the groups. Moreover, those transferees who were in work were not directly within
our field of interest since they had not lost their link with the labour market. The issue of
supplementing transfers with earnings is a separate topic. The share of these people in the
whole population is small.
Thirdly, the survey does not provide information on transfer entitlements prior to getting
a job in the case of people in work. Therefore, we dropped people who were currently in
work, as it was impossible to draw conclusions on the impact of obtaining transfers in the
past based on their present vocational activity.
Finally, we decided on a group of people who do not work at all, are of working age and
who live off cash transfers or support from others – this is the data base. Taking into account
the objective of the research, this population can be divided most effectively into 6 groups
of people:
1)Those getting disability pensions,
2)Those getting family pensions (except I-group distabled),
3)Registered unemployed receiving unemployment benefits,
4)Those getting other kinds of non-earned income (other than pensions and
unemployment benefits),
5)Registered unemployed maintained by others4,
6)Those maintained by others but not registered as unemployed (except all kinds of
students).
The first three groups (recipients of disability and family pensions, and unemployment
benefits) are transferees; the fourth group (getting other non-earned incomes) is not a pure
transferee group since it may include also people with income from property or gifts and
negative savings. Although the sum of such incomes in the National Accounts constitutes a
45
2 This is justified by the fact that in the 4th quarter of 2000, among the LFS population of women (men) at the effective
retirement age and above (55+/ 58+) most women (77%) and men (84%) were on retirement, or I/II group or farming
disability pension. In 2000, the official effective retirement age in Poland (in the employee system) was 55.9 for women and
58.9 for men.
3 From the definition I-group invalids are not able to work or even to exist alone. Indeed, only 3% among I-group disabled
had any work in the 4th quarter of 2000 survey. Therefore, by dropping this group of beneficiaries we made a negligible mistake
of not including potential labour suppliers.
4 'Other' represent here another member of a household getting his/her income from working in the non-agriculture or
agriculture sectors, or from non-earned source.

very small share in the gross income (below 4% in 2000), we should keep the structure of
these incomes in mind while interpreting the results. The other two groups are the
controlling groups with people who do not get any income and are financially supported
(maintained) by other household members. We separated the registered unemployed from
the total number of those maintained in order to test the difference in the attitude to work
between registered unemployed without benefit and registered unemployed entitled to
benefit. We also expect that among the maintained, those who are registered unemployed
have different objectives in their attitude to work than those who are not registered.
However, the potential advantages of being registered (like health insurance for all those
registered regardless of benefit entitlement) may somehow complicate this picture.
A direct investigation of how receiving transfers impacts finding a job could only be made
based on the panel data. The question would then be: which group of people is out of work
longest or how getting/losing entitlement to social transfers changes a person's approach to
looking for and taking work? The LFS scope and methodology hinder the panel analysis.
Firstly, questions in the survey do not refer to the person's past income and do not combine
this with data on his or her previous employment. This means that we do not know how
long the transferee has been receiving the transfer and what his/her simultaneous working
activity was. Secondly, the duration of the panel is too short. The same person is surveyed 4
times over 1.5 years. With some people getting unemployment benefit over a period of 1.5
years we are not able to observe sufficient numbers of people or to observe them over a
long enough period after the benefit expired.
Therefore, the analysis has to refer to the less direct relationship between social transfers
and their impact on acquiring work. The goal of the investigation will be to check how getting
a transfer is related to attitudes to taking a job.
In the LFS, we can identify three main questions which refer to the motivation to acquire
a job:
1) Are you seeking a job?
2) If you had ever worked in the past, how long have you been out of work?
3) If you claim you are seeking a job, what efforts have you made to find it: active
(searching via private offices, answering advertisements, distributing advertisements,
going directly to enterprises, trying to set up own business, other) or non-active (no
efforts, searching via friends or family, or registering at the public office, which we
consider to be a very minor effort)?
The answers to these questions by our 4 groups of transferees, as opposed to 2 groups of
those who are maintained, was the subject of analysis for 2 different points in time: 4th quarter
of 2000 in comparison with the 4th quarter of 1997 (see Profile analyses of the subject of
investigation in Section 2.2.2). The 4th quarter LFS was chosen because in 1994-1998 it was by
46

far the least subject to seasonal variations than other quarters. 2000 was the last available year
at the time of the empirical study. 1997 was chosen for the reference time as a contrasting
period to the end of 2000: in 1997 the Polish economy experienced its fastest growth during
the transition (6.8%), unemployment stood at the relatively low level of 10.3% and the
expansive social policy in 1996 should have manifested itself in wide access to transfers in 1997.
The last quarter of 2000 saw a substantial slowdown in economic growth (2.3% yoy), higher
unemployment (15.1%) and more restrictive policies towards social transfers.
2.2.2. Statistical Analyses of Work-Related Characteristics Among
Transferees and Others
Statistical analyses of the original LFS population
Before we concentrate on the relationship between social transfers and attitudes to work
within our sub-population, let us screen the original LFS population (before dropping certain
groups) according to the source of income. Knowledge about the relative weight of each
survey participant allows us to make the transition from the LFS population to the wider
Polish population. Numbers for the whole population are more interesting than numbers for
the LFS population because they can be related to other labour markets and social statistics.
Based on the 4th quarter 2000 LFS and GUS weights, we achieved a 30.6-million
population at or over the age of 15 (Table 2.6). Of all kinds of social transfers (11 million
entitlements all together), the biggest is the number of retirement pensions (5.4 million),
followed by the number of disability pensions (3.1 million); the smallest is the total of
unemployment benefits (0.5 million). These figures based on the survey correspond to the
numbers from the general statistics of pensioners (see Table 2.3). The number of people
who get support from others is 7.0 million, including 1.7 million maintained by recipients of
some form of social transfer or other non-earned income. The working population accounts
for 14.9 million people, with 80% involved in the non-agriculture sector.5
The possibility of combining social transfers with earned income from work (both in non-agriculture
as well as agriculture) is exercised moderately. Among transferees, the biggest share
of working people is in the group of disabled (15%), followed by the group of family pensioners
and recipients of other non-earned incomes (all together 12%); slightly smaller in the group of
retirement pensioners (9%), and the smallest share among the unemployed (3%).
Social transfers overlap between the five groups of transferees (D, F, U, O, P) to a very
small extent. For example, only 1% of family pensioners also receives retirement pensions;
in other groups only 1%-2% of transferees combine their transfers with other non-earned
incomes, which in some cases did not need to be social transfers.
47
5 Note: the definition of a working person is based here on the fact of having work in general as opposed to the LFS
definition where the working person has to actually work in the week of the survey.

Table 2.6.Population, 15+, according to main sources of income, at the total population level *, in
thousands of people, 4th quarter 2000.
Profile analyses of the subject of investigation
The total LFS 15+ population was reduced to individuals who are realistic providers of
labour and, at the same time, willing to be active on the labour market (intermediary sub-population).
This meant dropping retirement pensioners and all the other people above the
effective retirement age, the I-group disabled, as well as all kinds of students. Before we drop
48
Group of people
according to income
source **
D F U OMne Mna Ma Wna Wa P Total
D – getting disability
pension 2636 0 0 16 4 7 0 250 219 12 3143
F – getting family
pension 0 678 0 10 3 30 2 54 47 4 828
U – getting
unemployment
0 0 402 10 4 26 2 3 11 0 459
benefit
O – getting other
non-earned incomes 16 10 10 912 36 62 7 92 30 10 1185
Mne – maintained
by recipients of
4 3 4 36 1544 61 13 15 15 0 1694
non-earned income
Mna – maintained by
workers in non-agriculture
7 30 26 62 61 4368 26 53 32 4 4669
Ma – maintained by
workers in
agriculture
0 2 2 7 13 26 534 10 19 0 613
Wna – working in
non-agriculture 250 54 3 92 15 53 10 10586 491 256 11809
Wa – working in
agriculture 219 47 11 30 15 32 19 491 1954 229 3047
P – getting
retirement pension 12 4 0 10 0 4 0 256 229 4904 5418
TOTAL number of
income sources 3143 828 459 1185 1694 4669 613 11809 3047 5418 -
TOTAL number of
people - - - - - - - - - - 30640
With one source of
income 2636 678 402 912 1544 4368 534 10586 1954 4904 28518
With multiply source
of income - - - - - - - - - - 2122
Notes: * Weights of individuals in the LSF sample allow transforming the sample into the whole population; for the sake of simplicity,
in case of triple sources of one person's incomes he/she is considered only twice; the number of such persons is negligible as it accounts
only for 23 people at the population level;
** The table can be read from left to right or from top to bottom; each cell gives the number of recipients. Combining incomes
from two different sources is not excluded in the table; therefore some of the individuals are counted twice and that is why the total
number of income sources is higher than total number of people. People with a single source of income are represented on the diagonal.
Source: Own calculations based on LFS.

people who have worked and are therefore not of interest for us here, let us look at the
structure of the intermediary sub-population (Table 2.7).
Table 2.7. Intermediary sub-population's recipients of different kinds of income, at the total population
level, in thousands, 4th quarter 2000 *
The total number of potential providers of labour and, at the same time, those willing to work
is 18.8 million. The number of jobs in the non-agriculture and agriculture sectors (from Table 2.6)
represents 79% of this figure. So, if we assumed that all people were retired at the effective
retirement age, all retirement pensioners and I-group disabled were forbidden to work, and all
49
Group of people
according to the income
source
D F U OMne Mna Ma Wna Wa Total
D – getting disability
pension
1197 0 0 12 0 4 0 191 115 1520
F – getting family
pension
0 131 0 8 2 5 0 43 16 206
U – getting
unemployment benefit
0 0 397 9 4 26 2 3 11 453
O – getting other non-earned
incomes
12 8 9 801 23 40 5 84 27 1010
Mne – maintained by
recipients of non-earned
income
0 2 4 23 860 19 2 9 12 931
Mna – maintained by
workers in non-agriculture
4 5 26 40 19 1947 6 32 27 2107
Ma – maintained by
work in agriculture
0 0 2 5 2 6 180 6 10 211
Wna – working in non-agriculture
191 43 3 84 9 32 6 10204 481 11053
Wa – working in
agriculture
115 16 11 27 12 27 10 481 1864 2564
TOTAL number of
income sources 1520 206 453 1010 931 2107 211 11053 2564 -
TOTAL number of
people - - - - - - - - - 18776
With one source of
1197 131 397 801 860 1947 180 10204 1864 17580
income
With multiply source of
income
- - - - - - - - - 1197
Note: * All the comments for Table 2.6 also valid here.

students were not interested in taking jobs, the maximum unemployment rate (ceteris paribus)
would be 21%. At the end of 2000, the actual figure was somewhere below this level.
Within the group of potential and desired labour suppliers there are 3.2 million people
who receive transfers and other non-earned income (17% of the total); 2.7 million of these
are people who do not work at all (14% of the total). These ratios indicate how working
people are burdened by adults who receive social transfers. If one adds to this people who
are supported directly within the household (3.2 millions of the maintained) then the share
of non-working population is 34%. Despite that some of the maintained may be performing
certain work at home (e.g. housewives) the burden of non-working people in the group of
potential suppliers is quite high.
After excluding working groups and people with multiple incomes we obtain the final
sub-population. We must also reorganise the group of unemployed who are maintained by
others. We should also investigate separately those unemployed who registered at labour
offices (though without entitlement to benefit) and those who did not register. We believe
that the non-registered have quite different objectives in job seeking and should not be
analysed together. Table 2.8 presents the final sub-population broken down into 6 groups of
non-working population.
Table 2.8. Final sub-population's recipients according to their main source of income, 4th quarter 2000
The final sub-population consists of almost 8,400 people at the LFS sample level (which
is roughly 18% of the overall 4th quarter 2000 sample), or 5.4 million people at the
population level. The group of transferees is almost equal to the group of people maintained
50
Group of people according to their
main income source
LFS sample level Population level
structure,
% of total
thousand of
people
structure, %
of total
structure, %
of sub-total
D – getting disability pension 1 867 22.3 1 176 21.7 47.3
F – getting family pension 199 2.4 130 2.4 5.2
1 217 14.6 786 14.5 31.6
incomes
RU – registered unemployed
getting benefit
642 7.7 396 7.3 15.9
Transferee sub-total 3 925 47.0 2 487 45.8 100.0
RM – registered unemployed
2 484 29.7 1 593 29.3 54.2
maintained by others
NRM – non-registered unemployed
1 948 23.3 1 348 24.8 45.8
Maintained sub-total 4 432 53.0 2 940 54.2 100.0
TOTAL NUMBER OF PEOPLE 8 357 100.0 5 428 100.0 -

by others. This is a positive characteristic since these two groups will be mainly opposed to
each other.
Almost half of all the transferees receive disability pensions (47%); the smallest is the
group of family pensioners (5% of all transferees). Those who are maintained are almost
equally spread between registered and non-registered unemployed. Registered unemployed
with benefit entitlement are a much smaller group than registered without incomes. This is
a result of the substantial scale of structural unemployment, where only 20% of the
registered unemployed enjoy benefit entitlement.
Work-related characteristics of the final sub-population
Reasons of quitting/ending the job – for those who have worked previously
19% of the final sub-population have never worked. These are usually young people
(68% in the 15-25 age group), which corresponds with growing unemployment among
young graduates in Poland. If we look at maintained people versus transferees, a relatively
larger number of the maintained has never worked. This corresponds with the average age
of the maintained, which is 10 years lower than the age of the transferees (33 and 43 years
old respectively).
Among the population that has ever worked, the highest percentage (32%) lost their jobs
due to abolition of their work place. Since the beginning of the Polish transition public sector
employment has fallen systematically; in the late 1990s this fall process intensified (by 7-8%
annually6). The second most popular reason for quitting a job is retirement or other pension
entitlement (27% of these who have ever worked). 18% of respondents cited family reasons
for leaving their job, though there are very strong gender differences here – these are mostly
women who interrupt their working lives to bring up their children.
Poles seem to be relatively passive employees – only 3.8% (out of these who have
worked) left jobs due to unsatisfactory conditions at work (including financial conditions).
Most of them lost their jobs due to decisions made by employers (48%).
Within the groups of transferees and maintained people there are some obvious
correlations, for example quitting a job because of receiving a pension – in the case of
disability pensioners – or quitting a job for family reasons (mostly to bring up children) among
female 'other' transferees or maintained by others.
Period out of work
On average (measured by both the arithmetic mean and the median), among those who
have ever worked transferees remain unemployed slightly longer than maintained people (64
months versus 61 months, according to the mean, or 46 and 37 months, respectively,
according to the median). The type of a transfer is more important – the disabled stopped
51
6 GUS 2001.

working on average 90 months ago (7.5 years) and unemployment beneficiaries – 14 months
ago. Such discrepancies must result, to a large extent, from differences between the systems
in which transfers are granted. Disability pensions used to be assigned permanently, and if
assigned temporarily then were based on individual health conditions; in contrast to this, the
period of unemployment benefit assignment is restricted and quite short (up to 6 months in
most cases, and up to 18 months in the case of structural unemployment). A long period out
of work in the case of disablement may reflect the generous character of this system, but
also individual health conditions. However, if we exclude II-group disabled and disabled
farmers and leave only III-group of disabled, we do not get a far shorter period of being out
of work (84 months on average).
Seeking work
50% of the sub-population are not seeking work (are part of the non-active labour
force). Among transferees the ratio is even higher (66%), whilst for the maintained it is two
times lower (33%). Transferees appear relatively uninterested in getting work. Looking for
a job is again very highly differentiated between kinds of transfers. The least active are
disabled (91% do not seek work), while unemployment beneficiaries are the most active
(only 16% do not seek work). Among the maintained, the registered unemployed who do
not seek work are in a clear minority (13%, that is less than in the case of the unemployed
receiving unemployment benefits). As expected, the rest of the maintained represent a far
higher percentage of non-job-seekers (40%).
It is very interesting to juxtapose the above information with the working activities of the
transferees. Among the intermediary sub-population (realistic providers of labour and, at the
same time, desired to be active in the labour market) the share of the transferees or the
maintained which are active in the labour market is moderate (Table 2.9). In 2000,
transferees represent on average 42% activity: 15% through working and 26% through
seeking a job. 58% of transferees are passive on the labour market. This ratio can be seen
as another indicator of how transfers deter people from entering the labour market.
The least active are disability pensioners (close to 27% of the disabled). These, who are
active, mostly work rather than seek a job. A part of the disabled may not be active due to
health problems (despite this we suspect that a large number of the II-group disabled are
capable of working as their pensions were assigned in the generous system in the mid-
1990s). Also, disability pensions provide the strongest financial stability of all transfers and
this may well be another reason why some of the disabled are passive on the labour market.
More active are other transferees (41% for family pensioners and 48% for 'other'
beneficiaries). The registered unemployed are the most active of all, although those who are
maintained by others represent a higher rate of activity than these receiving a benefit (83%
and 77% respectively). This could suggest that the unemployed without benefits have
52

Table 2.9. Transferees or the maintained that are active on the labour market, 4th quarter 1997 and 2000
greater motivation to be active on the labour market. All the registered unemployed are
mostly active through seeking work, which indicates that unofficial jobs play a marginal role.
There is a clear difference between the registered and non-registered unemployed who are
maintained by others. The latter are more than two times less active on the labour market.
These people appear simply to have decided not to work. This may include women who
decided to continue bringing up their children after maternity benefits expired, or farming
family members which help in the farm while claiming to be non-working.
For pensioners (disability or family) labour market activity is presented much more via
working than is the case with benefit recipients (from the group of 'others' and unemployed
with benefits). This tendency may result solely from the fact that pensioners are much less
restricted from linking up pensions with earned incomes, which is why we observe much
smaller percentage of beneficiaries working. One could also try to explain this with the
complementary character of pensioners' work on top of pensions – they may tend to have
much smaller work requirements and so more easily find work.
When one compares labour market participation in 1997 and in 2000 one can see it is
higher in 2000, in particular in the case of maintained people and only insignificantly in the
case of transferees. The worsening of the economic situation did not really raise the
activity of transferees but rather of the maintained. We cannot, however, claim that the
transfers provided transferees with more stability than other household members did
towards those they maintained. The latter ones may also be more active on the labour
market for different reasons – e.g. younger, boom generations who have just started to
53
Group of people according to their
main income source
Share of people active
in the labour market
in % of total in the
given group
by working by seeking work
1997 2000 1997 2000 1997 2000
D – getting disability pension 33.6 26.9 26.4 20.1 7.2 6.8
F – getting family pension 28.2 41.4 18.8 28.6 9.4 12.8
incomes
30.0 48.1 9.7 11.0 20.3 37.1
RU – registered unemployed
getting benefit
90.0 76.9 7.0 3.1 83.0 73.9
Transferee sub-total 40.1 41.7 17.9 15.4 22.3 26.3
RM – registered unemployed
79.1 83.1 2.2 1.2 76.9 81.8
NRM – non-registered
unemployed maintained by
others
36.6 40.8 8.4 4.9 28.3 36.0
Maintained sub-total 53.1 63.2 6.0 3.0 47.1 60.3

enter the labour market. What changed over the analysed three year period was the
structure. Disabled worked less in 2000 than in 1997, and therefore their activity dropped.
The registered unemployed with benefit entitlement also lowered their labour activity, by
working but most of all by work seeking. It is not so clear why the unemployed were
seeking work less in 2000. One explanation could be that among the unemployed with
benefit the share of the long-term unemployed (who qualified for the benefit for the
subsequent time) increased. And the longer the person is unemployed – though receives
some funds – the less active he or she becomes despite the upcoming benefit expiry.
However, the share of the long-term unemployed in the total number of the registered
unemployed did not really change during that time (around 45%). Also, it seems that the
financial incentive to stay unemployed weakened: the ratio between the average level of
unemployment benefit and the average wage in the economy lowered from 29% in the
4th quarter of 1997 to 20% in 2000.
Recipients of family pensions and other non-earned income earners in contrast
increased their labour activity from around 30% to above 40%, although in a different way
– family pensioners managed to increase their employment, whilst 'others' intensified job
seeking. All the maintained – both registered and non-registered – slightly increased their
labour market involvement. This may reflect either financial stimulus (due to a general
worsening of the economic situation) or some demographic changes, as well as some
other structural issues.
However, the average activity of transferees may be considered high. This may
indicate that transfers did not push people into complete passiveness. High labour activity
may also suggest that the level of the transfers was low enough to motivate them to find
alternative sources of incomes. However, other interpretations of these figures also make
sense. It may be that most of those who are active on the labour market either tried to
milk the social security system and get transfers as additional income sources on top of
their work. It is also possible that because they have been seeking work for a long time
they were forced to get some pensions or benefits. It is impossible to verify the causality
of this relationship.
Immediate employment
Returning to the final 2000 sub-population (that is, transferees and the maintained who
do not have any work), among those who seek work, the willingness to take on a job
immediately (this or next week) is relatively high and similar for the transferees and the
maintained (93% of all who seek work). The level of differentiation between the groups is
very small. The most desperate to take on a job are the registered unemployed without
benefit (94.4%); only slightly less eager are the unemployed with benefit (94.0%). In other
groups around 90% of people seeking work want to take on a job immediately.
54

Minimal salary required by the unemployed willing to take on a job immediately
The average net wage in the economy in the 4th quarter of 2000 was around 1391 PLN7.
The minimum wage required by those who are willing to take on a job immediately is
someway below the average (841 PLN). One could expect that for people who are able to
mix earned incomes with transfers (i.e., many disability and family pensioners as well as some
recipients of other non-earned incomes) the desired wages would be a supplement to their
transfers and therefore its level should be lower. However, the differences in respect to the
minimum wage between the groups are small: the lowest minimum required wage is for
disability pensioners (753 PLN), the highest for unemployment beneficiaries (898 PLN).
Small differences may indicate that social transfers are low enough to motivate people to
work for almost as much as people who would not have any alternative incomes. A low level
of the minimum expected salary in the group of registered unemployed may suggest either
strong desperation to find a job (after remaining out of work for a long time), or the low
working status connected to poor education or experience.
Efforts undertaken to find a job (by the unemployed willing to take a job immediately)
The structure of answers is similar for all groups of transferees and those that are maintained.
Most of them contain themselves to rather passive forms of job searching. Close to 40% of people
seeking work are registered with public offices, another 23% asked relatives about job
opportunities. More active were these who approached employers directly or answered
advertisements (33% in total). A relatively small number of people did not make any efforts (1-2%).
Concluding, the total activity of transferees in the labour market is quite high, although lower
than the activity of the maintained. From this we can deduce that social transfers are not necessarily
associated with significant labour market passiveness. When we consider only people without
work, two times more people do not seek work among transferees than among the maintained.
The average period out of work matters substantially across the groups. However, an important
factor would be here the duration of the transfer entitlement as well as some family characteristics
of the maintained. Other work-related characteristics do not point to strong differences between
the groups. Therefore, out of all potential characteristics representing the attitude to acquiring
work, 'seeking work or not' was chosen for the further econometric analysis.
2.2.3. Econometric Analysis of the Relationship between Transfers and Labour Supply
The relationship between getting a transfer and not seeking work has come up both in
the statistical analysis as well as in statistically significant correlations. However, as many
55
7 Own calculations.

studies on the labour market mechanisms prove, labour supply depends on a range of factors
simultaneously. Simple two-variable correlations do not take into account interrelations
between the explanatory variables. The fact of getting, e.g., maternity benefit may have a
different impact on work attitude in the case of a single woman who, at the same time, plays
the role of family head, and different in the case of a married woman whose husband is self-employed.
Taking into account information on marital status, the main income of the
household and the status in the family, we should come up with a better estimation of the
maternity benefit impact on seeking work.
What we are interested in is not only the simultaneous relationships between more than
two variables, but also comparing the strength of the association between them. This
requires constructing a regression model. Since we believe that transfers have an impact on
attitudes to work,8 the explanatory variables will be the kind of transfer or financial support
and other individual characteristics, and the dependent variable will be 'seeking work'.
The list of available explanatory variables includes:
1) source of income,
2) gender,
3) marital status,
4) position in the household,
5) number of adults in the household,
6) number of children in the household,
7) age,
8) educational level,
9) main source of income in the household,
10) place of habitation (size of town or rural area),
11) region (voivodship),
12) duration of a period without work.
Theoretical background
The dependent variable in the model is dichotomous, taking a 1 value when the person
seeks work and 0 value when he/she does not. Therefore we use the logit regression. The
explanatory variables have the form of both dichotomy and quantitative variables.
The main idea behind the logit model is to find the relationship between the probability
(Pi) that Y – the dependent variable – will take a 1 value (given the characteristics of an i-individual
Xi, that is, a combination of explanatory variables Xij), and the characteristics of an
i-individual.
Pi = E(Y=1 (Xi) = β1 + β2 Xi )
56
8 Opposite direction causality is also possible: unwillingness to work or inability to find work may create the demand for
transfers.

Besides some other statistical requirements, in order to receive a non-linear
relationship between Pi and Xi, as well as to be sure that the probability Pi lies between
0 and 1, the equation Pi has been modified and the new representation (logit model)
looks as follows:
Li = ln (Pi / (1-Pi) = β1 + β2Xi),
where (Pi / (1-Pi) is the odds ratio in favour of Y=1. The higher the odds ratio the bigger
probability that Y=1.
It can be shown (using calculus) that
dPi/dXi = β2 Pi/(1-Pi),
dPi = β2 Pi/(1-Pi) dXi
It "shows that the rate of change in probability with respect to X involves not only β2 but
also the level of probability from which the change is measured"9.
The slope coefficients βj
tell us the direction and relative intensity of how the incremental
change in X impacts the change in the probability that Y=1 (given the initial probability of Y=1).
Outcomes
The outcome of the model is presented in Table 2.10.1 (for 2000) and Table 2.10.2 (for 1997).
The model offers interesting conclusions about the relationship between transfer
entitlements and work seeking as well as other individual characteristics. The main question
asked in the study is answered as might be expected: people receiving transfers are less likely
to seek work than these who do not have any source of income and are maintained by
others. The disabled are the least likely to seek work, followed by family pensioners and
'other' beneficiaries. The non-registered unemployed maintained by others do not show a far
greater interest in seeking work than all other beneficiaries. The most active in seeking work
are the registered unemployed without benefit entitlement. They are even more active than
unemployed who receive benefit. The effect of the statistical analysis that unemployment
benefits may not motivate labour activity (Table 2.9) was also confirmed by the model, that
is, after taking into account all other characteristics of the individuals. The fact that the higher
probability of seeking work was assigned to the maintained groups suggests that, despite
how easy support from household members may be, the maintained may have more
motivation to earn their own income (maintained people on average are younger than the
transferees by 10 years – aged 32 as opposed to 42).
57
9 Gujarati (1995) p. 559.

58
Explanatory variables Coefficient Significance Stand. dev.
Main income (ref.: Maintaining the registered
unemployed without benefit)
Disability pension
Family pension
Other non-earned income
Unemployment benefit
Maintaining the non-registered unemployed
Male (ref.: Female)
Marital status (ref.: Married)
Single
Widowed
Divorced
Marital status of women (ref.: Married woman)
Single woman
Widowed woman
Divorced woman
Head of family (ref.: Other position in the family)
Number of adults 15+ in the household
Number of children under 15 in the household
Age
Education level (ref.: Basic vocational)
Pre-primary and primary
Secondary
Vocational secondary
Post-secondary
Higher
Main source of household income (ref.: Employment)
Farm ownership or farm use
Self-employment
Pension
-4.311
-3.531
-2.572
-0.639
-2.035
1.322
-0.072
0.306
0.401
1.129
1.384
0.780
-0.063
0.078
-0.053
-0.028
-0.304
0.078
0.188
-0.155
-0.564
-0.684
-0.543
-0.058
0.858
0.824
0.000
0.000
0.000
0.000
0.000
0.000
0.616
0.504
0.104
0.000
0.006
0.012
0.533
0.008
0.114
0.000
0.000
0.554
0.034
0.493
0.008
0.001
0.000
0.515
0.000
0.000
0.131
0.240
0.118
0.142
0.096
0.101
0.144
0.458
0.246
0.206
0.501
0.310
0.101
0.029
0.034
0.004
0.087
0.133
0.089
0.226
0.212
0.213
0.144
0.089
0.208
0.130
Table 2.10.1. Outcome of the model, 4th quarter 2000
Dependent variable: seeking work=1, not seeking work=0

59
Table 2.11.1. Outcome of the model, 4th quarter 2000 – cont.
0.308
0.322
0.351
0.364
-0.102
-0.213
-0.382
-0.440
-0.290
-0.101
-0.125
-0.352
-0.340
-0.411
-0.347
-0.098
-0.250
-0.012
-0.113
-0.002
2.424
0.014
0.001
0.005
0.000
0.550
0.192
0.019
0.010
0.089
0.550
0.492
0.042
0.042
0.035
0.062
0.574
0.160
0.955
0.551
0.000
0.000
0.126
0.096
0.125
0.093
0.171
0.163
0.163
0.170
0.171
0.169
0.181
0.174
0.167
0.195
0.186
0.175
0.178
0.219
0.189
0.001
0.246
Town size (ref.: Rural area)
< 10,000 inhabitants
10,000-50,000 inhabitants
> 100,000 inhabitants
Region (ref.:) Dolnooel¹skie1
Mazowieckie
OEl¹skie
Wielkopolskie
Pomorskie
Zachodniopomorskie
£ódzkie
Lubuskie
Ma³opolskie
Kujawsko-pomorskie
Opolskie
OEwiêtokrzyskie
Warmiñsko-mazurskie
Podkarpackie
Podlaskie
Lubelskie
Period out of work
Constant
Log likelihood
Pseudo R² (Cox & Snell R Square)
Pseudo R² (Nagelkerke R Square)
Sample size
5867
0.407
0.543
6803
Notes: * Reference variables are those which equal 1 for the biggest number of cases;
** The final size of the sample included in the regression is smaller than the original size presented in Table 2.8 because
including the variable 'period out of work' missed cases of people who never worked. However, it did not really change the
specification of the model.
10 Voivodship ordered according to GDP per capita in 2000, from the highest to lowest; Dolnooel¹skie – the reference, is
on the fourth position.

Another important characteristic confirmed by the model is the impact of the out of
work period on attitudes to work. The correlation is negative, which means that the longer
the person stays out of work the less likely he/she is to seek work. This is in line with the
general belief in the labour economy that returning to work becomes the more difficult, the
longer the period of job-lessness. However, the conclusion that transfers provide a good
enough alternative to earned income would be too simplistic. There may be other factors
(not represented in the model) responsible for the loosening of the link with the labour
market.
In the case of men, the probability of seeking work is higher than for women. This is not
surprising as women in Poland still play to a large extent a traditional role in a family, whereas
men still tend to be the main provider – in the sample, men are heads of families two times
more often than women are. The negative correlation between being a head of a family and
seeking work would mean that the declarations of heading the family are traditional, while
income responsibility in the family does not stick to this traditional model any longer.
However, the coefficient by this variable is insignificant.
Marital status is not really a statistically significant factor in attitudes to work. The only
significant factor – at 10% – is a divorced status. The coefficient indicates that divorced
people are more eager to work than married people. However, when we test the
interaction between seeking work and the marital status of women we see an opposite
situation: married women are least eager, and divorced women have the higher
probability of looking for a job. The probability is even higher for single and widowed
women. This is most likely the effect of women's role in the family, which provides less
motivation to work. On the other hand, single women may be forced to seek work due
for financial reasons.
Size of family is clearly an important factor. The probability of seeking work is positively
correlated with the number of adults in the household (the bigger the family the higher the
probability). On the other hand, the bigger the number of children up to 15 years old, the
smaller probability of looking for a job. Despite the fact that this coefficient is significant at a
low level (11%), a possible explanation could be that small children (we do not know their
age structure) require more attention and care, whilst older children require more financial
support from their parents. Therefore, we might find that the coefficient would be positive
for parents of older children.
The probability that a person will look for a job decreases with age. This may result from
civilisation differences between generations, as Poland continues to undergo not only
economic but also social transition. Younger generations seem to become more open to the
emerging dynamics of the labour market and are more able to respond positively to it. Also,
the system of transfers may be still considered rather generous, and access to transfers
naturally increases with age.
60

The statistical significance of different education levels is uneven: pre- and primary,
vocational secondary and higher education are significant, with vocational level as a
reference. Most active in work seeking are people with vocational secondary, followed by
those with vocational education. People who stopped at primary or pre-primary education
show the least motivation to seek work. This suggests that vocational and secondary
vocational education tend to induce labour activity among the unemployed. Most
surprisingly, however, is that people with university degrees have the lowest probability of
looking for a job. One might expect them to be more motivated to seek work (in search of
self-fulfilment, higher income requirements, pressure of the environment) than people with
lower education.
Individuals from households where the main income comes from self-employment are
less eager to seek work than in the case of employment income. This may be related to the
higher on average incomes of self-employed. Individuals from a household where the main
income is from farm ownership or farm use are least eager to seek work. Here the
explanation could hardly be high incomes. It is rather the poor geographical mobility of
farmers or involvement in family farming which constrains these people from seeking work.
But if the main income in the household is from unemployment benefit or 'other' non-earned
incomes, individuals seem to have more incentives to look for a job. Here, again, the reason
may refer to the financial situation of the family. Pensions as the main income in the
household do not explain the probability of seeking work (statistical insignificance), that is,
other factors must play a role.
The place of habitation of unemployed people has a significant impact on their job
seeking. Probability of seeking work is generally higher in cities than in rural areas, and the
bigger the town the higher the probability. Labour mobility as well as wider opportunities to
find work in bigger towns may have an important influence on the relationship.
There is no clear correlation between the probability of seeking work and the region
where a person lives. With Dolnooel¹skie voivodship as a reference, the correlation is
statistically significant in the case of only 7 voivodships (out of 16). The first possible
conclusion is that, if we neglect Opolskie and Pomorskie, the higher the unemployment rate
in the voivodship, the higher probability that a person will seek work (Table 2.11). This
would mean that the unemployment situation within each voivodship is relatively uniform.
Another conclusion could be that the higher GDP per capita, the higher probability that
people seek work. Such a relationship could mean that people in richer regions have a
greater chance of finding work or at least approaching their unemployment with more
optimism and creativity. However, if we reverse the causality, the relationship could mean
that regions with people more active in seeking work are simply more prosperous.
Nevertheless, the territorial structure with 16 voivodships is general and there must be
plenty of other factors responsible for differences in job seeking across regions.
61

Table 2.11. Voivodship as an explanatory variable in 'work seeking' – statistically significant relationships
The correlations between seeking work and available individual characteristics of non-working
people and the outcomes of the model are quite similar for the 1997 and 2000
populations.
Different groups of transferees and maintained people present similar probabilities of
seeking work: the least probability is in the case of disability pensioners, followed by family
pensioners and recipients of other non-earned incomes. The non-registered unemployed
who are maintained by other household members have only a slightly higher probability of
seeking work than the previous three groups (the disabled, family pensioners and 'others').
Again, the registered unemployed that do not receive benefit are the most willing to seek
work – more than the unemployed who still enjoy benefit entitlement.
Gender mattered in 1997 to the same degree as in 2000, where unemployed men were
much more likely to seek work than women. Although marital status as such is insignificant
here, the interaction of gender shows that single or divorced women are more active in job
seeking than married women. The family structure has a similar impact in 2000. Again, the
more adults in the family the more motivations to search for work. On the other hand,
children in the family are a negative factor in 1997 too. Being the head in the family also does
not really seem to matter.
The influence of age differences in the labour market appear to have increased over time,
as the age factor is less significant in 1997 than in 2000. However, the relationship between
age and job seeking is negative in both years (the elderly are less eager to seek work).
Secondary education increases the probability that a person will seek work – both
secondary vocational and simple secondary (insignificant in 2000) have higher coefficients for
1997 than vocational education. Primary education has no impact on job seeking. However,
as in 2000 a university degree also provides less motivation to seek work.
Tendencies with regards to the main income source in the household are also quite similar
in both years. However, income from farming was statistically insignificant in 1997 and turned to
be significant in 2000. This could indicate that the factors making rural people reluctant to seek
62
Voyevodship,
ordered acc. to the coefficient
(from the highest to lowest)
GDP per capita
in 2000,
thousand PLN
Unemployment rate,
end of 2000,
%
Location
Dolnooel¹skie 16.3 18.4 South-West
Zachodniopomorskie 15.9 20.8 North-West
Kujawsko-pomorskie 14.1 19.2 North
OEwiêtokrzyskie 12.4 16.6 Central-East
Ma³opolskie 14.2 12.2 South
Wielkopolskie 16.7 12.5 Central-West
Opolskie 13.3 15.7 South-West
Pomorskie 16.1 16.6 North

63
Main income (ref.: Maintaining the registered
unemployed without benefit)
Disability pension
Family pension
Maintaining the non-registered unemployed
Male (ref.: Female)
Marital status (ref.: Married)
Single
Widowed
Divorced
Marital status of women (ref.: Married woman)
Single woman
Widowed woman
Divorced woman
Head of the family (ref.: Other position in the family)
Number of adults 15+ in the household
Age
Education level (ref.: Basic vocational)
Pre-primary and primary
Secondary
Vocational secondary
Post-secondary
Higher
Self-employment
Pension
-4.776
-3.381
-2.359
-0.760
-2.236
1.279
0.140
-0.037
-0.079
0.874
0.444
0.885
0.141
0.125
-0.094
-0.007
0.000
0.318
0.303
-0.051
-0.563
-0.240
-0.500
0.088
0.335
0.856
0.000
0.000
0.000
0.000
0.000
0.000
0.350
0.938
0.769
0.000
0.420
0.009
0.185
0.000
0.006
0.155
0.999
0.024
0.001
0.829
0.028
0.309
0.001
0.354
0.062
0.000
0.149
0.293
0.137
0.125
0.103
0.111
0.150
0.467
0.270
0.208
0.550
0.341
0.107
0.031
0.034
0.005
0.091
0.141
0.093
0.238
0.256
0.236
0.153
0.095
0.179
0.158
Table 2.10.2. Outcome of the model, 4th quarter 1997

64
Table 2.10.2. Outcome of the model, 4th quarter 1997 – cont.
Explanatory variables – cont. Coefficient Significance Stand. dev.
Town size (ref.: Rural area)
< 10,000 inhabitants
> 100,000 inhabitants
Region (ref.:) Dolnooel¹skie15
Mazowieckie
OEl¹skie
Wielkopolskie
Pomorskie
Zachodniopomorskie
£ódzkie
Lubuskie
Ma³opolskie
Kujawsko-pomorskie
Opolskie
OEwiêtokrzyskie
Podkarpackie
Podlaskie
Lubelskie
Period out of work
Constant
0.201
0.113
0.348
0.253
0.050
0.140
0.501
0.449
0.070
0.076
0.276
0.016
0.020
0.231
0.134
0.113
0.357
0.417
0.087
0.006
1.512
0.201
0.113
0.348
0.253
0.741
0.360
0.003
0.010
0.699
0.650
0.223
0.927
0.914
0.298
0.563
0.561
0.056
0.075
0.660
0.000
0.000
0.146
0.098
0.126
0.098
0.152
0.153
0.168
0.175
0.182
0.167
0.227
0.170
0.181
0.222
0.231
0.195
0.187
0.234
0.197
0.001
0.256
Log likelihood
Pseudo R² (Cox & Snell R Square)
Pseudo R² (Nagelkerke R Square)
Sample size
5341
0.403
0.548
6541
Note: * Reference variables are those which equal 1 for the biggest number of cases.
11 Voivodships could not be ordered in 1997 according to GDP per capita because the new territorial structure was
introduced only in 1999 and all the macroeconomic indicators before 1999 are presented in the old territorial breakdown.

work (e.g. low mobility), have become more relevant over time. Another interesting factor is
that the unemployed from households where the main incomes come from unemployment
benefit were less eager to seek work than 'other' transferees. In 2000, their activity in work
seeking was relatively higher. This, combined with data showing that the level of unemployment
benefit compared with the average wage worsened, would suggest that low benefits motivate
more unemployed family members to look for work. The importance of the financial status of
the family also appears to manifest itself, as in 2000, in a lower probability of seeking work by
individuals from families with self-employed incomes as the main source than with employee
wages; 1997 the wages were on average lower than the earnings from self-employment.
Where people live was an insignificant factor in 1997. The conclusion from this could be
that the differences in attitude to work activity increased over time as the economic
slowdown in the bigger cities outpaced that in small cities or rural areas.
The influence of the regional factor cannot be matched fully with the economic
characteristics of regions, as the economic indicators for 1997 are available only in the old
voivodship structure. In 1997, only 4 voivodships were important factors in job seeking, and
only 2 of them were also significant in 2000. People from these 4 voivodships had a lower
probability of seeking work than in Dolnooel¹skie, though the lowest probability was in
relatively affluent voivodships (Wielkopolskie and Pomorskie). Two poor regions on the east
(Podlaskie i Podkarpackie) are in the middle.
In the 1997 model, as in the 2000 model, the longer a person stays out of work the less
eager he/she is to seek work.
2.3. Concluding Remarks
The analysis – both statistical and econometric – shows that the relationship between
seeking work and receiving transfers should not be considered in a simplistic way. There are
great differences among transferees in their willingness to take work. However, obtaining or
not obtaining transfers partially explains attitudes to work. Most of the available individual
characteristics matter as well. Additionally, we believe that the list of these characteristics
misses some very important variables, such as health, the number of small children in the
family (we know only the total number of kids up to 15 years old), the general level of
generosity of the social security and the protection system, e.g. the average period of
entitlement for different kinds of transfers (we know that on average disability pensions
provide the greatest stability as they are assigned usually for much longer than other
transfers; followed by family pensions, 'other' transfers and unemployment benefits, though
individual information on the entitlement duration is not available).
65

The outcomes of the statistical and econometric analysis, however, provide a lot of
valuable suggestions. In all the tests the job-less transferees with most stable transfers have
a less eager attitude to work. However, all groups of transferees exhibit relatively high labour
activity (either through working or job seeking). This also supports the thesis that the level
of transfers is generally low and that transferees tend to look for complementary sources of
income. This would mean, furthermore, that social transfers do not provide disincentives to
work and force transferees to earn additional amounts rather than to look for a permanent
job. This also means that all transferees that are passive on the labour market for other than
financial reasons must be receiving relatively low incomes. These people may be pushed
towards poverty. The model shows that people from households where the main source of
income is from unemployment benefit or other benefits are most eager to seek work.
Moreover, the average duration of their time out of work is relatively long (over a year in the
case of the unemployed with benefit, and several years in the case of other transferees).
The causality of the relationship between seeking work and receiving social transfers
cannot be clearly identified. In every case an opposite interpretation is equally possible:
people who cannot find work due to a worsening economic situation and rigid labour market
were forced by poverty to seek social transfers. The question "do these transfers help people
to regain motivation to look for a job or not?" becomes relevant.
The vast majority of transfers in Poland are pensions. Once they are assigned they cannot be
taken back unless there is a change in the law (e.g. an overall reform of the system). Therefore,
social and labour policies can be practically addressed really only to the access to benefits (which
have a limited and shorter duration) and the future assignments of all kinds of pensions.
Taking into account the relatively high labour activity on the part of transferees,
macroeconomic policies – including social and labour market policies – should concentrate
above all on creating opportunities on the labour market to absorb labour resources.
Secondly, the objective should be to sharpen the eligibility criteria for transfers as well as
practises of assignments in order to improve transfers targeting but also to define a level of
transfers that would be adequate to the modest but real costs of living. Only realistic level of
benefits can prevent transferees from falling into poverty.
66

Introduction
During the political and economic transformation in Poland after 1989 the environment in
which Polish households operated changed dramatically. There was a marked rise in the
number of cases of social pathology, which fast became a publicly discussed topic. Phenomena
that had earlier been almost unknown, such as poverty and unemployment, became social
problems almost overnight. Experts pinpoint alcohol as a key factor in the outbreak of
pathological behaviour (Sieros³awski, Szumlicz 1995). However, during the transition limited
research was conducted into social exclusion connected with alcohol abuse. The research
presented in this chapter focuses on social assistance recipients and looks specifically at the
relationship between alcohol abuse within households and their social and employment
characteristics. Available administrative data from social assistance centres allow quantitative
analysis of the relationship. They also enable us to identify the characteristics of those
households affected by poverty and with officially reported alcohol problems.
The first section provides a description of the pattern of alcohol consumption in Poland.
Definitional problems of alcohol abuse, alcohol dependency and their relationship with social
and economic factors are described in the second section. The following analysis concentrates
not on alcohol consumption for the whole population, but on cases of reported alcohol abuse
by recipients of social assistance. Profile analysis shows the structure of social assistance
recipients, and compares this to the situation of those recipients that have reported alcohol
problems in their households. The logit model is used to investigate the interrelations
between reported alcohol problems, family background and economic situation of social
assistance recipients (section nine). Conclusions outline some of the policy implications for
targeting socially marginalised groups due to poverty and alcohol-related problems.
67
Part 3
Alcohol Abuse and Poverty
Agnieszka Sowa*
* I would like to express my thanks to Prof. Stanis³awa Golinowska for her professional assistance, and to Miriam Beblo
and Charlotte Lauer for their technical advice.

3.1. Alcohol Abuse as a Social Problem in Poland
Public opinion tends to see alcohol dependency as one of the most important problems
afflicting Polish society during the transformation process. According to a public opinion poll
taken by the PARPA in the summer of 1998 on a sample of citizens over the age of 18,
alcoholism is rated the fourth most important problem affecting the country (Sieros³awski,
1998). Alcoholism1 (9.9% of responses) lagged behind violence on the streets (20%),
unemployment (18.4%) and falling living standards (10.2%). Respondents were also asked
which problems were the most important for their local community. In this context,
alcoholism was seen as the second most important problem (15.1% of respondents), after
unemployment (22.7%).
Over the post-war period alcohol consumption patterns in Poland were similar to those
in most Western European countries. The first 30 post-war years were characterised by an
increase in alcohol consumption. According to official estimates, average alcohol
consumption in Poland measured in liters of pure alcohol consumed per capita2 in the 1950s
was almost double the figure for the pre-war period. It increased further in the 1970s
(Moskalewicz, Sieros³awski, Szymanowski, OEwi¹tkiewicz, Zieliñski 1991).
68
Table 3.1. Average consumption of alcohol in Poland per capita, 1938-1990
Years
Average alcohol consumption
(in litres of pure alcohol per capita)
1938 1.5
1950-1954 3.1
1955-1959 3.7
1960-1964 3.9
1965-1969 4.7
1970-1974 5.9
1975-1979 7.8
1980-1984 6.7
1985-1989 6.8
1990 6.0
Source: Moskalewicz J., Sieros³awski J., Szymanowski T., OEwi¹tkiewicz G., Zieliñski A., 1991.
1 In the survey the question was asked of "alcoholism" not "alcohol dependency" or "alcohol-related problems", as the term
"alcoholism" is easily recognized by the public and popularly understood as excessive drinking violating social norms and
resulting in social dysfunctions.
2 This measure is commonly acceptable and allows for comparisons between countries, however one has to remember
that this is the average measure, while real consumption varies significantly between groups, with strong concentrations of
consumption – according to estimations half of alcohol consumed in Poland is consumed by 10% of the population.

The increase in alcohol consumption can largely be explained in terms of changes in the
type of consumption. In 1938 the most common alcoholic beverage was spirits, while in the
1960s the share of spirits dropped while wine and beer increased. However, the
consumption of wine and beer did not entirely replace the consumption of vodka. The share
of vodka remained at 70%, the highest share for any European country. Alcohol
consumption dropped in 1981 due to the social and economic crisis, as well as restrictions
on the supply of alcohol during Martial Law. But later in the 1980s alcohol consumption
increased once again.
The 1990s decade brought more changes in the pattern of alcohol consumption. Official
data from 1990 show a drop in alcohol consumption, though the reliability of such figures can
be questioned due to the widespread increase in access to illegal, smuggled alcohol,
especially spirits.
Figure 3.1. Average consumption of alcohol in Poland, in litres per capita, 1992-2000
8
7
6
5
4
3
2
1
0
1992 1993 1994 1995 1996 1997 1998 1999 2000
According to GUS and PARPA data, alcohol consumption in Poland at the end of the 1990s
stabilised at 6.73-6.75 litres of pure alcohol per capita per annum. However, this data point to
general trends and thus doesn't offer fuller information on the level of alcohol consumption:
1. Data on alcohol consumption compiled by GUS apply only to registered alcohol trade
and do not take into account illegal trade. PARPA estimates illegal production and
smuggling totals between 20% to 30% of total legal trade. This means that unrecorded
consumption may be at around 2 litres of pure alcohol per capita per annum.
2. Data given by GUS relate to the whole population, including children, while some
other estimates (e.g. by WHO) relate only to adults (15 as the threshold). According
69
3. Alcohol Abuse and Poverty
Source: GUS/PARPA 2001.
Spirits
Wine
Be er
Tota l

to WHO, actual average consumption is about 1.5-2 litres of pure alcohol per capita
per annum higher.
When taking these factors into account average alcohol consumption per adult (aged
15+) would be 3.5-4 litres of pure alcohol higher than GUS estimations (that is 10-11 litres
of pure alcohol per adult).
During the 1990s drinking patterns changed, with the consumption of spirits falling and
beer consumption increasing. However, high frequency of drinking combined with heavy
drinking (consumption of large amounts of alcohol on one occasion) rates remained more or
less the same. This type of drinking is perceived as related to a high risk of alcoholism.
PARPA estimates that 5% to 7% of the population (2-2.5 million people) has a problem
with alcohol and about 2% of the population (800,000 people) is alcohol dependent.
Another 4% (1.5 million people) are adults living with alcohol dependants, and 4% of the
population are children who live in families with at least one alcohol dependant.
3.1.1. Alcohol Abuse, Alcohol Dependency and Alcohol Related Problems
– Towards a Definition of the Phenomenon
The consumption of alcohol has always played a part in societies' cultural, social and
economic affairs. Aside from any political or religious guises it may have adopted over the
years, alcohol's negative side is visible as quickly, if not more so, as any positive effects it
may have.
The functions and dysfunctions of alcohol use can be analysed from different perspectives
depending on the aspects of the phenomenon they emphasise. Social pathology theory, social
disorganization and deviant behavior, as well as public health approaches are among the most
influential sociological and psychological theories.
Social pathology theory emphasises the behaviour of individuals and focuses on
phenomena that do not meet commonly accepted social rules. Pathology is analysed at the
individual, rather than the social, level, without analysis of social context. Individuals whose
behaviour is inconsistent with social rules are seen as "sick" and alcohol abuse or alcohol
dependency is treated as a "sickness". American psychiatrist Jellinek formulated the
following definition of alcoholism as "any use of alcoholic beverages that causes any
damage to the individual or society or both". Alcoholism is treated as a permanent,
irreversible and progressive sickness. The theory emphasises the loss of control over
drinking habits and inability to stop drinking. Treatment via isolation and drying-out in
special centres is based on this concept. Its weakness is that it treats the phenomenon in
the isolation from any social context.
70

Social disorganisation theory seeks reasons for social problems in the social system and
relates them to the process of social change. Various phenomenon, including pathological
drinking, occur during periods of rapid social change and breakdown of moral standards, the
theory suggests. A similar approach was proposed by Merton. His deviation theory argues
that the behaviour imposed by social change leads to the composition of new strategies.
These strategies can be either innovative/constructive, or destructive. Drinking is one way
of adjusting to a changing social environment.
The above-mentioned theories, combined with epidemiological data on the quantities of
alcohol that can be safely consumed without causing negative health effects,3 provide the
background for traditional definitions of alcohol abuse and alcohol dependency. In 1952 the
WHO's definition of alcohol abuse was an excessive drinking that may lead to disturbance of
an individual's social functions at the workplace, school or home; such as absenteeism at the
workplace, lower productivity, absenteeism at school and negligence in performing
household duties (Schafer 1996, PARPA 2001). Excessive drinking may also lead to
involvement in gambling, or aggressive behaviour and law-breaking.
Alcohol abuse has to be distinguished from alcohol dependency. They differ in
quantitative and qualitative terms, although the relationship between the two categories is
relatively fluid. Alcohol abuse may lead to alcohol dependency, but not necessarily (Schafer
1996, PARPA 2001). Alcohol dependency is a multi-dimensional sickness that damages
individual psychological, social and mental functions. It is characterised by excessive drinking,
combined with the compulsion to drink and loss of control. Dependants tend to drink
without respect to their obligations or health. Alcohol dependency also causes psychological
dysfunctions (Mellibruda 1999):
– destructive self-image: low self-respect and negative perception of life, suicidal
thoughts, feeling of guilt and shame;
– destructive relations with others: aggressive behaviour, isolation from social
environment (family etc.), antisocial behaviour;
– breakdown of moral standards.
These symptoms lead to a deficit in social skills and social functions that are not easily
reversible. They also demand time-consuming treatment long after quitting drinking.
The WHO's most recent comprehensive approach to alcohol dependency targets
problems that often go hand-in-hand with alcohol use. This public health approach
concentrates not only on those who are alcohol dependent, but also on those for whom
alcohol might lead to individual or social problems. The policy concentrates on early
recognition of the problem, early intervention and rehabilitation. The goal of intervention is
71
3 According to WHO standards, the amounts of alcohol that can be consumed without the risk of future health damages
are 20 grams of pure alcohol for women and 40 grams for men, up to 5 times a week;

the reduction of alcohol-related mortality and social and economic problems. Alcohol
related problems are distinguished from other phenomena in terms of the causality between
the problem (e.g. family dysfunction, loss of job, car accidents, suicides) and alcohol
consumption (Rehm 2001). This is not one-way cause and effect relation, as alcohol is one
of many different risk factors that may lead to dysfunctional behaviour. Alcohol related
problems may be incidental or long-term. For public health approaches, alcohol dependency
as defined as one of several alcohol-related problems. The approach targets the whole
drinking population for whom any negative effects of drinking are visible.
3.1.2.The Individual and Social Consequences of Alcohol Consumption
Alcohol consumption has harmful consequences.
Categorising some incidents as alcohol related implies that (Rehm 2001):
– an incident is related to alcohol by "chance", which means coexisting in time and place,
without setting a type of relation between the two, or
– alcohol is a risk factor in sparking an incident where there is an assumed – but
unproven – causal relation between it and the influence of alcohol, or
– there is a causal relation between alcohol and an incident, but some occurrences may
be partially due to alcohol, while others are direct consequences of alcohol.
The effects of alcohol are individual (often medical) or related to social behaviour.
Epidemiologists concentrate on the individual health consequences of alcohol consumption
– sicknesses or deprivation that alcohol directly causes or for which it is a crucial risk factor.
Recent analyses of the effects of alcohol include social problems related to alcohol. As
J. Rehm (2001) notes: "the social consequences of alcohol are changes, subjectively or
objectively attributed or attributable to alcohol, occurring in individual social behaviour or in
social interaction or in the social environment". Changes related to alcohol can be positively
or negatively evaluated by the individual and his/her environment (subjective assessment).
They can be objectively confirmed by, for example, alcohol tests.
The health consequences of alcohol consumption
Long-term drinking or heavy drinking increases the risk of depression and mortality risks.
Some illnesses are directly attributable to alcohol, such as alcohol dependence and alcohol
psychosis. In Poland, the share of people suffering from alcohol dependence has been rising
steadily, from 1.1 to 4.3 per 100,000 people between 1980 and 1995 (WHO 2001). The risk
of illnesses, such as chronic liver disease and cirrhosis is increased by long-term alcohol
consumption. Alcohol contributes to psychosis and personality disorganisation that in some
72

cases results in suicides. According to police records, in 1993 25% of suicides were
committed by intoxicated people (Sieros³awski, Szumlicz 1995). Intoxication is also one of
the major causes of car accidents, many of which end in death, permanent illness or disability.
Most of the health consequences of alcohol consumption affect an individual's situation
via sickness or disability, for example. However, alcohol also has social consequences (in
terms of the above definition). Treatment of alcohol related illnesses increases health care
system costs. Depression related to alcohol impacts on the social relations of the ill person,
his/her family life, and his/her activity on the labour market and productivity.
Social environment, friends and public safety
Alcohol consumption is strongly related to social situation. The peer group plays an
important role in the initiation of alcohol use, largely conditions the frequency and quantity of
further alcohol consumption and the probability of delinquent behaviour. Family and peer
associations also condition the age of initiation and act as models for alcohol use. Drinking
combined with risk-taking behaviour among adolescents may be a means of gaining status in
the group. Alcohol abuse in the social context is characterised by loss of psychological and
social barriers and the focus on the "here-and-now" appears to provide opportunities for
enjoyment, meeting new people, romance and sexual relations. The end results of alcohol
abuse and this loosening of constraints can often be violent behaviour, sexual assault, and
vandalism. The frequency of alcohol related violence among groups of friends and
acquaintances is hard to estimate as it is rarely reported. Few people report violent
occurrences within such groups to the police, while the police tend to disregard violence within
groups of alcohol abusers (Pernanen 2001). Alcohol consumption may also lead to behaviour
that threatens public safety: noisy behaviour in public places, increased risk of car accidents,
crime and violence. The frequency of alcohol related violent behaviour varies between
countries depending on legislation and social acceptance of alcohol and type of consumption.
In Poland, alcohol is seen as one of the major factors underlying criminal behaviour. According
to police statistics, about 25% of crime is alcohol related (Sieros³awski, Szumlicz 1995).
Family consequences of alcohol consumption
Alcohol related problems affect the abuser's family, spouses and children, above all.
According to estimates, for each person who abuses alcohol at least one other person close
to him/her suffers from his/her drinking (Maffli 2001). The consequences of being close to
drinking problems are of three types:
– Abusive or other forms of unacceptable behaviour towards relatives by the person under
the influence of alcohol. Such behaviour may be extremely aggressive, taking the form of
physical violence or sexual abuse. According to PARPA, 2/3 of children and 2/3 of adults
living in households with a drinking person have been subjected to domestic violence.
73

– Suffering social and economic problems, such as isolation from the community, lack of
money for basic family needs and household support due to incomes going on alcohol,
loss of employment, etc.
– Decomposition of family leading to family break-up and divorce.
While conceptually distinct, these consequences usually coexist. Another possible
consequence of alcohol abuse is the "contamination" effect on relatives: which often refers
to relatives exhibiting drinking tendencies.
Influence of alcohol problem on children
Children are most susceptible to alcohol related problems simply because they can do
little (if anything) to protect themselves. One risk is being subjection to prenatal alcohol
exposure. The most serious damage is foetal alcohol syndrome (FAT) characterised by
damage to the child's nervous system, growth defects and facial deformity. Other "alcohol-related
birth defects" are characterised by learning and behavioural disorders (Maffli 2001).
Children growing up in families with disorders due to alcohol often suffer from physical
violence, sexual abuse, abandonment, and emotional abuse. Research on parent-children
relations in families with alcohol-related problems show that children in these families tend
to be endangered by the risk of being drawn into drinking activities themselves. No causality
has been properly outlined in this case, and the risk is differently estimated by researchers,
but the relation is no doubt a significant one.
Children in families with alcohol-related problems are also more inclined than their peers
to take responsibilities inappropriate for their age. They often feel obliged to look after their
drinking parent(s), are involved in parental disputes and arguments, and are worried about
household finances, etc.
Influence of alcohol problem on spouses
The most evident influence of alcohol problem on the spouse is violence and marital
disruption. Evidence suggests that alcohol is not in itself a necessary or sufficient condition
for domestic violence, but that there is a strong positive association between the two. There
are also health and psychological consequences of alcohol and domestic violence. Alcohol
abuse as the cause of marital disputes may often result in marital break-up. Drinking is not
the only reason of divorce – there may also be financial difficulties, problems keeping down
a job, unpredictable behaviour, aggressiveness and violence towards partner and/or children,
all resulting from drinking. In Poland in the 1980s and 1990s alcohol abuse was cited as the
second most frequent reason for divorce – about 25%-29% of marriage break-ups were
caused by alcohol related problems (Sieros³awski, Szumlicz 1995).
74

Economic consequences of alcohol consumption
Alcohol consumption has economic consequences in several dimensions:
– social transfers from the budget to families requiring help due to problems resulting
from alcohol,
– impact on the education of children and adolescents, which further influences their
economic performance,
– direct influence on an individual's economic performance: absenteeism at work,
accidents resulting in death or injury, low productivity, theft, crime and inappropriate
behaviour, poor co-worker relations and low morale,
– costs for employer due to above listed poor performance of alcohol abusers,
– however, at the same time, companies producing alcohol are an important employer,
– indirect taxes imposed on alcohol consumption contribute to the state budget.
The list of economic consequences of alcohol consumption can be broadened, however,
for the purpose of this research the first two are taken as the most important.
The poor school performance of a child or adolescent may result from his/her direct
involvement in alcohol-related activity or an alcohol problem within the family: prenatal
exposure to alcohol or alcohol abuse by parents. Prenatal exposure to alcohol affects a child's
problems with concentration, hyperactivity, and academic difficulties. Research confirms that
living in a family with alcohol problems is associated with children's' poor performance at
school and truancy.
On the other hand, many adolescents themselves also engage in drinking activity that may
influence school attainment. An adolescent's involvement in alcohol consumption may also be
partially attributable to lack of parental care and/or being surrounded by peers who abuse
alcohol. These factors indicate a vicious circle of drinking, truancy and poor school performance.
This cycle leads often to low participation in the labour market (Rehm, Rossow 2001).
Alcohol abuse by adults is a risk factor for poor performance at work due to absenteeism,
drinking at the workplace or low productivity caused by hangovers, and is associated with
unemployment. Here causality is not clear and can work in both directions: heavy drinking
may lead to unemployment, or loss of work may cause drinking. Some research studies also
suggest that using alcohol may be related to taking up jobs for shorter periods of time than
non-drinkers do.
Finally, while analysing the economic costs of alcohol consumption one has to include the
problem of poverty. Lower activity of alcohol abusers combined with high expenditure on
alcohol results in the household's lower economic status. In 1993, on average, expenditures
on alcohol amounted up to 20%-25% of total household expenditures on food
(Sieros³awski, Szumlicz 1995). However, these expenditures are strongly differentiated
among households. According to PARPA, in 1993 for about 4% of Polish households
75

(480,000 households) money spent on alcohol counted for 50% or more of their total food
expenditures, and 30% spent on alcohol more than 10% of their total food expenditure
(Sieros³awski, Szumlicz 1995). For these households alcohol expenditures are a significant
burden to their budgets.
3.2. Empirical Study
3.2.1. Research Hypothesis
This study examines social assistance recipients suffering from poverty and various social
pathologies, including alcohol abuse. The research seeks to answer the question: what is the
likelihood of alcohol becoming a factor in falling into poverty; poverty measured in terms of
the entitlement to social assistance benefits. The aim of the research is to uncover the
individual and household characteristics that may predispose recipients of social assistance to
household-based alcohol-related problems. The central research hypothesis suggests that
alcohol abuse is connected with the labour market situation, which in turn is related to family
dysfunctionality, especially family decomposition, in households receiving social assistance
benefits. There is no necessary direct corelation between poverty and alcohol related
problems, or causality can work in both directions: alcohol abuse can be a consequence of
poor labour market situation and poverty, or poverty can cause high level alcohol
consumption. A subsidiary research hypothesis is that alcohol has a negative impact on an
individual's employability, worsens the situation of the household, acts as a barrier to
improving the situations of households and as such is a factor of social exclusion. The
coexistence of economic and family dysfunctions related to alcohol and poverty causes a
vicious circle of social exclusion.
3.2.2. Data Source
In order to target alcohol problems and social exclusion patterns the research is based on
administrative data for poverty and dysfunctionality. Analysis is restricted to those people who
are dependent on assistance from a variety of social assistance programmes provided by Social
Assistance Centres (SAC). The database used is POMOST. It was launched by the World Bank
Promotion and Services Project (1999) and is managed by the Ministry of Labour and Social
Assistance. The database contains information on social benefits paid to SAC recipients. The
goal of the system is to provide the Ministry of Labour with the possibility of monitoring social
76

assistance and improving social benefits targeting. In the long run the system is aimed at
improving the quality of services provided to social assistance beneficiaries as well as
rationalising and controlling spending. Despite the system still not being fully in operation
nation-wide, there is a representative sample of 15% of all SACs (375 out of 2556).
The database is structured in terms of information on each social benefit given to social
assistance recipients in each quarter of the year starting from the first quarter of 2000. To
understand the database's construction it is important to be acquainted with the procedure
of granting social assistance. SACs provide beneficiaries with different types of benefits
depending on the household's situation. Each household's situation is evaluated by a social
worker, who decides also on the appropriate type of benefit. Each adult household member
can apply to the SAC for assistance. Social assistance, defined by the Social Assistance Act
(1990), is addressed to the family, though benefits are granted to individuals. Therefore, each
household member may be granted a benefit in cash or in kind individually depending on
his/her needs, taking into account the situation in the entire family. Any individual who applies
for benefits has to take a "family interview", which determines any assignation of benefits.
Since the benefit is formally addressed to a household the "family interview" not only
concerns the individual situation of a person but also the situation of each household
member. Social benefits are subjected to an income test and dysfunctionality criteria,
according to which a person is entitled to the benefit if the household's income is below the
threshold level and, at the same time, the household faces one of the problems listed by the
Social Assistance Act (Golinowska 2002):
– poverty,
– parentlessness,
– homelessness,
– single maternity,
– unemployment,
– physical or mental disability,
– long-term illness,
– problems in running a household, especially in single–parent families or multi-children
families,
– alcoholism4 or drug addiction,
– difficulties in adjustment after imprisonment,
– ecological disaster.
While the Social Assistance Act lists 11 dysfunctions that provide the basis for
determining whether an applicant receives social benefits, in the "family interview" these
become 13 problems. Alcoholism and drug addiction are split into two separate categories,
77
4 The Social Assistance Act uses the term "alcoholism" and the same form is used in the family interview; alcoholism is
understood here as an alcohol abuse or alcohol dependency that lead to household social dysfunctions.

and one category (the incidents) is added. Any individual who takes the "family interview"
may pinpoint several problems that his/her household may suffer from. The dysfunction
might concern either an individual (or several individuals) within the household e.g.
unemployment, alcoholism, or the whole household e.g. poverty, ecological disaster. Since
the individual who takes the interview gives information on the whole household's situation
it is not clear which household member suffers from which problem (e.g. which household
member suffers from alcohol abuse or alcohol dependency).
The social worker decides which benefit should be given to different individuals within
the household entitled for social assistance. Each benefit is recorded in the database
separately. Each individual receiving some kind of a benefit is given his/her own identification
number. It may also the case that not all members of a household that is entitled to social
assistance are listed in the database, but only those individuals who are directly entitled to
some kind of benefit (including children). Each beneficiary can also potentially receive several
benefits during one quarter; in the database there can be more than one observation per
person. Identification numbers allow a tracking of individuals within each quarter as well as
over quarters and perhaps years. For each individual the following information is entered:
type of problem within the individual's household, individual's gender and age, family
composition, city size, labour market situation, income per capita within the household,
individual's income source, household's maintenance source.
The analysis herein was based on this database covering all four quarters of 2000 for a
sample of 375 SACs, covering 216 233 social assistance recipients.
3.2.3. Methodology of the Analysis
The analysis adopts a quantitative approach to alcohol related problems among social
assistance recipients. Administrative data from SACs were used for the first time for such
analysis and therefore needed detailed preparation for the purpose of the research.
Quarterly data were aggregated into a one-year database, and the structure of the database
was changed from data organised by benefit to data organised by individual. We identified a
group of recipients who indicated alcoholism as one of the problems in their household at
least once during the whole of 2000 (0-1 variable, where 0 – household does not suffer from
alcohol problem, 1 – household suffers from alcohol problem). Other variables describing
the individual's or household's situation were recoded for the purposes of the research.
To explain the interrelations between alcohol abuse among social assistance beneficiaries
and its socio-economical determinants we conducted a profile descriptive analysis on two
groups: a general social recipients sample and the group of recipients that reported at least
once that their household suffered from alcohol abuse. In further analysis we investigated the
78

strength of the interrelationships between alcohol problems in the household and the
recipients' socio-economic situation using a logit model. The aim of the analysis was to come
up with the probability that alcohol abuse is a cause of households applying for social
assistance, depending on the individual/household situation.
3.2.4. Characteristics of the Studied Groups
Social assistance beneficiaries
Demographic and social characteristics
There is a large disproportion between male and female social assistance recipients, with
women accounting for 66.9% of total social assistance beneficiaries in 2000, against 33.1%
for men. This disproportion can be explained by several different factors. Partly it is a
consequence of targeting some benefits to single mothers, as single maternity is one of the
dysfunctionality criteria outlined in the Social Assistance Act. The high proportion of women
among social assistance beneficiaries can also be explained by the role of maternity in Polish
households, where women take responsibility for providing a part of household income,
especially if the mail partners are unemployed.
In terms of age, most social assistance beneficiaries fall into one of two groups: children
under the age of 18 and individuals between the age of 35-44.
Table 3.2. Social assistance recipients by age
This disproportionality in the age of social assistance beneficiaries is coherent with other
research on poverty in Poland, which depict – surprisingly – that it is not the elderly that are
most threatened by poverty but families with children (World Bank 1995). Children who are
79
Age groups Frequency Percent
17 years old and less 45 069 20.9
18-24 27 364 12.7
25-34 39 741 18.4
35-44 45 004 20.8
45-54 31 566 14.6
55-64 9 347 4.3
65 years old and more 17 851 8.3
Total 215 942 100.0
Missings 291 0.1
Source: Own calculations.

recorded as "beneficiaries" of social assistance generaly do not receive benefits in cash, but
95% of them are granted "childrens' meals at school or kindergarden", 2.7% are granted
"family benefits", and each other type of benefits are granted to less than 1% of children.
According to information from the Ministry of Labour, meals for children at school are the
most commonly given type of benefit, almost in every case of families with children that are
entitled to social assistance. However it may be the case that some children in households
receiving social benefits are not entitled (individually) to any kind of social assistance,
including meals at school. Such children are not included in the database.
Close to 50% of the social assistance beneficiaries' households are married couples with
children. This is in line with earlier research on poverty in Poland (World Bank 1995), which
found that children as a group are significantly affected by poverty. The second group
threatened by poverty are one-person-households. This could indicate that these
households are mostly threatened by social exclusion and a high level of alcohol
consumption. Among social assistance beneficiaries the number of single mothers is quite
high. This tends to result from their underprivileged economic situation as well as due to the
targeting of social assistance to them (and that is why it is widely represented in the
database).
Table 3.3. Structure of social assistance recipients by family composition
As also earlier research shows (Golinowska 1997), the highest share of recipients of social
assistance is among farmers. In a sample half of the families live in the countryside. Poverty in
rural areas is connected with high levels of long-term unemployment and low real incomes
80
Family composition Frequency Percent
Marriage with 1-3 children 57 899 30.7
Marriage with 4 and more children 35 863 19.0
One-person-household 32 758 17.4
Single mother 30 669 16.3
Multi-person household 10 157 5.4
Multi-family household 7 689 4.1
Marriage without children 5 780 3.1
Concubinage (with or without children) 5 061 2.7
Single father 2 326 1.2
Supplementary family 334 0.2
Total 188 536 100.0
Missings 27 697 12.8

from agricultural production. Such results appear to indicate that larger cities have suffered
less during the economic transformation than farmers and inhabitants of small towns.
Table 3.4. Social assistance recipients by place of living
Economic characteristics
Social assistance benefits are income tested. An average income per capita in recipients'
households is significantly lower than the social assistance income threshold. In 2000 these
were PLN 168.7 and PLN 350 respectively.
Most beneficiaries are not active on the labour market as they are either unemployed or
they are applying for social assistance help due to permanent illness or physical disability. It
is necessary to state that the system of social assistance is targeted at individuals and families
suffering from illnesses and permanent disabilities either themselves or family members.
Among social assistance recipients there is higher share of people passive on the labour
market (51.2%) than the registered unemployed (39.2%) and the working poor (9.2%).
Moving from an individual to the household level analysis, the maintenance income of a
household is non-earned for 65% of household-recipients (with at least one family member
entitled to social assistance).
Social assistance beneficiaries with alcohol-related problems
An alcohol problem in the Social Assistance Act is outlined as one of the dysfunctionality
criteria entitling individuals to apply for social benefits. Research concentrates on individuals
suffering from alcohol problems within their closest social environment who admitted in the
interview to the existence of an alcohol problem in their household. However there are
several problems related to alcohol, such as unemployment, low income and family
decomposition, that entitle households to apply for social assistance. Alcohol problems are
81
Place of living Frequency Percent
Rural areas 118 783 55.8
Town of up to 10 thousand 13 515 6.3
Town of 10-20 thousand 16 511 7.8
Town of more than 200 thousand 3 771 1.8
Total 213 003 100.0
Missings 3 230 1.5

Table 3.5. Social assistance of household-recipient by main maintenance sources
often only shamefully admitted to. Therefore, some individuals who suffer also from other
problems that qualify them for social assistance – often alcohol related (e.g. unemployment)
– may decline from admitting alcohol abuse as a problem in their household. This means that
analysing the situation of individuals who admit alcohol related problems allows a targeting
of the core of the issue, that is, the population for whom alcohol problems are evident and
hard to hide.
This analysis examines the relationship between alcohol abuse/dependency and socio –
economic factors at the individual and the household level. For this purpose, a variable
representing admitting alcohol abuse was created. The variable helped to aggregate a
population that – at least once – pinpointed alcohol problems as one of the reasons of
applying for social assistance. This does not mean that social assistance benefits were given
to all of them because alcohol abuse/dependency was the households' main dysfunction. In
2000, alcohol abuse was admitted by only 1.9% of all social assistance beneficiaries.
Demographic and social characteristics of social assistance recipients reporting alcohol
related problems
More men than women report having alcohol problems. This is particularly striking if one
takes into account the fact that social assistance recipients are overwhelmingly women. The
figure for men reporting alcohol dependency is almost double that for women. This supports
the findings of earlier research into alcohol consumption across society as a whole, and not
only among the poor. According to a 1989 survey, men consumed 4.5 times more alcohol
than women (Moskalewicz, Sieros³awski, Zieliñski 1991).
The same research points to age as the second most important factor in determining
alcohol consumption. Its findings suggested that the highest frequency of alcohol
82
Household’s main maintenance source Frequency Percent
Non-earned incomes other than pensions 56 012 37.5
Employment 39 592 26.5
Old-age, family or disability pension 38 411 25.7
Agriculture 13 324 8.9
Self-employment 1 989 1.3
Total 149 328 100.0
Missings5 66 905 30.9
5 There is a high level (30.9% of total social assistance recipients) of missing information on the household's main
maintenance sources which may be due to lack of willingness on the part of recipients to provide information on their
maintenance (which could also be from working in the shadow economy).

consumption is found among persons between 40-49 and that the lowest is found among
young (between 14-19) and older people (over 60). Analysis of alcohol consumption among
social assistance recipients conducted 11 years later partially confirms these results. The
highest frequency of reporting alcohol related problems is among social assistance recipients
between 35-54 (2.5%-3.0% of all 35-54 years olds), while those between 18-24 rarely
report having an alcohol problem (0.4%). It is interesting that children were often reported
as suffering from alcohol related problems (2.5%). This indicates that among poor
households a significant number of children live in families with dysfunctions and poor
economic situations, often related to alcohol.
Reporting an alcohol problem within a household is significantly differentiated in terms of
family composition. Single mothers and families with up to 3 children least frequently report
alcohol related problems despite these types of families being highly represented in the sample.
Alcohol abuse is reported relatively rarely by what we may call average families (defined here
as a married couple with up to 3 children). Alcohol abuse is reported most often by social
assistance recipients living in concubinage with or without children (3% of all the
concubinages), by single persons (2.9%) and single fathers' households. Alcohol problems are
reported relatively frequently also by larger families with more than 4 children (2.5%).
Summing up, the relationships between reported alcohol problems in the family and gender
and/or family composition is strong. The share of households with an alcohol abuse dysfunction
is above average among one-person-households. These are more often headed by men than
women; 33% of men on social assistance and 12% of women live in single-households.
The last significant socio-demographic variable in outlining the pattern of alcohol abuse
among social assistance recipients is where people live. Although more than half of those served
by SACs live in rural areas, the highest share of reported alcohol related problems is people who
live in small towns with up to 20,000 inhabitants. Of these, the most threatened by alcohol
dependency tend to be individuals living in towns of between 10,000 and 20,000 (3.6%).
The economic situation of social assistance recipients reporting alcohol related problems
An individual's labour market situation has a significant bearing on the chances that he or
she will report having alcohol related problems. Those in work are more likely to report
having alcohol related problems (3%) than non-working social assistance beneficiaries or the
unemployed (2.5%) or those simply passive on the labour market (1.9%). In examining this
relation one needs to focus on labour market factors, including maintenance sources and
levels of income for families.
Alcohol related problems are reported most frequently by households' on
unemployment benefit and income from agriculture.
The relationship between household income per capita and reported alcohol abuse is
strong, in fact almost fully linear. The frequency of reported alcohol abuse is highest among
83

households with the lowest per capita income and decreases with higher income. Average
household per capita income in households with alcohol related problems is significantly
lower than average income per capita among social assistance beneficiaries in the sample.
Figure 3.2. Reporting alcohol related problems by household per capita income
3.5
3
2.5
2
1.5
1
0.5
0
To 50 51 - 100 101 - 150 151 - 200 201 - 250 251 - 300 301 - 350 351 - 400 401 - 450 451-500
The rough profile of those reporting alcohol problems can be formulated as follows: a
man, living in a single household, often working but not regularly, earning a very low monthly
income (up to 50 zlotys). At the same time, some social assistance recipient groups rarely
report having alcohol problems, for example single mothers.
3.2.5. Factors of Reporting Alcohol-Related Problems by Social Assistance
Beneficiaries – Outcomes of the Research
In order to explore the interrelationships between the reporting of alcohol problems and
other social indicators we ran a binary logistic regression, which allowed us to gauge more
systematically the relationships between reporting alcohol problems within the household
and the household's social and economic situation, identified in the profile analysis. It also
enabled us to exclude any spurious relationships.
The research goal was to determine the probability that social assistance recipients
would report alcohol problems. Its outcome should indicate the risk factors of reporting an
alcohol problem in the household. However, one should note that the relationship between
reporting an alcohol problem and the household's situation are not instrumental in any causal
sense and the direction of causality is not always clear. For example, it is impossible to
84
Household per capita income (PLN)
Frequency of repoirtng an alcohol problem

unambiguously state whether having an alcohol problem in the household results from
unemployment in the family, or that unemployment was the reason for alcohol abuse.
The results of the model should provide answers to the following hypothesis which was
formulated on the basis of the profile analysis:
1. Men more often report having alcohol problems than women.
2. Habitation (size of town/city) is an important risk factor in reporting household alcohol
problems; living in rural areas or a small town is positively correlated with reporting
alcohol abuse.
3. Reporting alcohol abuse is correlated with family composition – single people and
those living in broken families are more prone to report cases of alcohol abuse.
4. Unemployment of a family member and the poverty caused by it are significantly
related to alcohol abuse.
Explanatory variables are made up by various indicators of the household situation: the
social assistance recipient's gender and age, where they live, their type of a family, their
labour market position, the household's main source of income and household's per capita
income. Each of the variables is recoded into the set of binary variables. The reference
category is always the category with the highest frequency. The analysis was conducted on a
sample of social assistance recipients over 22 years old, or over 18 and running single
households. Children or adolescents living with their parents and reporting an alcohol
problem are excluded from the analysis since we assumed that in most cases they
themselves are not alcohol abusers and someone in their household is reported as such.
Poles tend not to leave home or live on their own until they are in their twenties.
While interpreting the results of the regression one should underline that the data
concern individuals who live in households with alcohol related problems. This does not
imply that the individuals themselves are alcohol abusers (although this might also be the
case). However, for the policy agenda it is important to note that individuals are strongly
affected by alcohol related problems in their closest environment (that is, households in
which alcohol abuse either by themselves or by other family member(s) leads to poverty and
social exclusion of the entire household). It implies, further, that receiving social assistance
due to having an alcohol problem is an indicator of the degree to which the household is
socially excluded.
The outcome of the model confirms some of the formulated hypotheses.
Gender is an insignificant factor in social assistance beneficiaries reporting alcohol
problems, contrary to expectations. However, men who are active on the labour market
(either employed or searching for a job) are significantly more likely to report alcohol
problems in the household. There seems to be an important relationship between
employment and the reporting of alcohol problems, though this has not yet been carefully
85

Table 3.6. Explanatory variables
studied (it is not known, for example, whether this applies to full-time, part-time or
temporary employment). As such, one cannot rule out the possibility that this relationship
may be spurious. The most interesting positive correlation is between male unemployment
and the reporting of an alcohol problem with a female social assistance beneficiary as the
reference. There is the additional effect of being unemployed and male. The coexistence of
these two features significantly increases the risk of living in a household that has reported
having an alcohol problem. The likelihood of an unemployed male reporting an alcohol
problem is higher than for an employed male.
In the regression, as in the profile analysis, age is also an important factor in reporting
own or family members' alcohol related problems. People between 18-34 are less likely to
report alcohol problems than those between 35-44. Although the results relate only to
households receiving social assistance, they tend to confirm the general trends in Poland,
with the highest levels of alcohol consumption in the group between 35-44 year old and the
lowest among 18-25 year olds (Sieros³awski, Szumlicz 1995).
Moving from individual to household features we find a positive and significant
86
1. Gender Male
Female
2. Age 18 – 24
25 – 34
35 – 44
over 45
3. Family type Marriage without children
Marriage with up to 3 children
Marriage with 4 and more children
Single mother
Single father
Multi-family/multi-person household*
Concubinage
Single household
4. Place of habitation Rural area
Town up to 20,000 inhabitants
20,000 – 100,000 inhabitants
More than 100,000 inhabitants
5. Individual labour market position Passive
Unemployed
Employed
6. Household’s maintenance sources Pension
Other than pension non-earned income
Employment
Agriculture
7. Income Per capita income in household
Note: * A few families or a few persons not in a marital relationship living in one household.

87
Table 3.7. Logistic regression results
Dependent variable: reporting alcohol problem within the social assistance beneficiary's household as one
of the reasons for applying for social assistance, values: 1 - yes / 0 - no
Explanatory variables Coefficient Standard error Odds ratio
Individual features
Gender
Female
Male
Reference
-0.090 0.097 0.914
Labour market position
Passive
Employed
Unemployed
Reference
0.943
0.136
***
**
0.085
0.071
2.569
1.145
Gender and employment
Female
Employed Male
Unemployed Male
Reference
0.147
0.453
***
***
0.195
0.110
Age
35 – 44
18 – 24
25 – 34
over 45
Reference
-1.722
-0.779
0.350
***
***
0.204
0.066
0.054
0.179
0.459
1.035
Household’s features
City size:
Rural area
<20,000 inhabitants
20,000 – 100,000 inhabitants
>100,000 inhabitants
Reference
0.382
0.082
-0.419
***
***
0.061
0.056
0.143
1.465
1.085
0.658
Family composition
Marriage with 1-3 children
Marriage without children
Marriage with 4 and more children
Single mother
Single father
One-person household
Multi person/family household
Concubinage
Reference
0.136
0.025
-0.971
0.063
0.592
-0.033
0.406
***
***
***
0.137
0.070
0.095
0.106
0.076
0.092
0.116
1.146
1.025
0.379
1.066
1.807
0.967
1.501
Household’s maintenance sources
Other than pension non-earned
income
Employment
Agriculture
Pension
Reference
-0.290
-0.283
-0.191
***
***
***
0.074
0.104
0.073
0.748
0.754
0.826
Per capita income -0.002 *** 0.001 0.998
Log likelihood
18215.489
Cox & Snell R Square
0.014
Nagelkerke R Square
0.070
Note: Level of statistical significance of the variables: * 10 percent, ** 5 percent and *** 1 percent.
0.494

1.503

relationship between living in small towns and reporting alcohol related problems. Social
assistance beneficiaries living in towns of up to 20,000 are more likely to report alcohol
problems within their households than social assistance beneficiaries living in rural areas. At
the same time, those living in towns of more than 100,000 are the least likely to report to
SACs that any member of their household has alcohol related problems.
The relationship between family breakdown and the reporting of alcohol problems is
significant only in the case of single mothers, one-person households, and concubinages, with
the average family (married couple, with up to 3 children) as the reference. Single-mother
households tend to be least prone to report alcohol problems, even more than the average
family. Single households are the most likely to report alcohol related problems. This may
imply that individuals living in these households are not only reporting alcohol problems but
most likely also drinking themselves. The other group that is only slightly less likely to report
an alcohol problem is households based on concubinage.
The outcome for the relationship between household maintenance sources and alcohol
related problems suggests that lack of employment among the economically active is the
most important factor in inducing the reporting of alcohol abuse among social assistance
beneficiaries. Households with incomes from some kind of economic activity or from
pensions are less likely to report alcohol related problems than those who live off non-earned
income.
3.3. Conclusions
The above analysis allows us to identify those groups of social assistance beneficiaries that
are most likely to have alcohol problems, either themselves or in their families, and to report
this to SACs. One of the factors most likely to induce reporting of an alcohol problem in
one's own household is unemployment of a household member, especially men. Individuals,
in turn, become less employable when living in households with alcohol related problems
and this is also a barrier to their reintegration into society. Without touching on the causality
of the relationship between alcohol problems and unemployment, reintegration of this group
into society and increasing their employability is extremely difficult. Those who are socially
excluded due to unemployment, poverty and alcohol dependency tend to be at the same
time least motivated to take up any activity that will enhance their position. They require
complex treatment: psychological assistance, increasing life motivation and social skills as
well as training to raise their labour market abilities. The social assistance addressed to such
people cannot focus solely on material help, but should also provide these beneficiaries and
their families with preventive actions and medical treatment. As the research has showed,
88

those who tend to report having alcohol problems most often live in single households, with
less support from spouses and therefore are in greater need of external motivation.
At the same time the research confirms the differences in deprivation between rural and
urban populations in Poland. Rural dwellers more often than urban dwellers apply for social
assistance. Moreover, alcohol problems are reported more frequently in small towns than in
rural areas, though this relationship might be spurious. It is possible that in rural areas high
alcohol consumption is not seen as a problem and therefore tends not to be reported.
However, despite the probable underreporting of alcohol problems in rural areas, alcohol
consumption is still higher in rural than in urban areas.
Overall, research based on official, administrative data from SACs does not tell the whole
story of alcohol related problems among the poor. What it does do is highlight some
important relationships between alcohol problems and employability, as well as other social
features of households that reports alcohol dependency problems. It provides a useful
framework within which policy implications can be formulated. The research also
complements previous studies, mostly qualitative analyses, on alcohol dependency and social
exclusion in Poland.
89

Introduction
The transition process from a centralised to a market economy in Poland has led to a
higher average standard of living within the population. However, transition has also been
accompanied by a deepening of inequality across households in terms of socio-economic
status. As economic differences between households in terms of their labour market access,
educational background, region and other socio-demographic characteristics are suspected
to prohibit the equality of socio-economic opportunities for children, this inequality might
become even more severe for the next generation. Indeed, the family represents the crucial
link that passes socio-economic endowments of the older generation (parents) to the
younger one (children). In particular, if parental poverty has a large impact on the educational
prospects of the children, poverty is likely to get passed on over generations because a low
level of education dramatically increases the risk of experiencing poverty in the sequel. The
transition process may exhibit a detrimental dynamic by which poor education and poor
opportunities all over the life-cycle perpetuate across generations. The most disadvantaged
families may be caught in a poverty trap, which may give rise to societal disruptions in the
medium run.
The purpose of this study is therefore to assess the extent to which parental poverty
is transmitted to the next generation via the educational attainment of children in Poland.
More precisely, we aim at identifying the link between an individual's socio-economic
background as mainly provided by the family and his/her educational attainment
throughout the transition period. Although the incidence of poverty in Poland as well as
the effects of state transfers to relieve poverty have been extensively investigated on a
91
Part 4
Family Background and Children's
Educational Attainment During
Transition
Miriam Beblo and Charlotte Lauer

general level by, for example, Golinowska (1996, 1997), Kotowska (1997, 1998), Szulc
(1997) and the World Bank (1994, 1995), a microeconometric analysis of the relationship
between family background and children's educational attainment during transition is still
missing. Available data sources such as the Labour Force Survey (LFS) provide
information on the social situation as well as on the education level of all household
members over the 1990s and therefore enable us to identify the characteristics of
disadvantaged families and to investigate how these relate to the educational outcomes
of children in such families.
In the remainder of the chapter, we will first give a brief overview of the extent and
structure of poverty in Poland. Then, we will outline the mechanisms by which poverty may
be transmitted from parents to children through the educational achievement of children. In
the third section, we will present our empirical analysis of the determinants of children's
education using data from LFS. After a description of the Polish education system, we
provide an overview of the structure and the developments in enrolment rates and
educational levels for our sample of LFS. Next, we apply an ordered probit model to
investigate the relationship between family background and children's educational attainment
on a multivariate basis and to identify in particular the role of the parents' financial or labour
market situation. The chapter concludes with a discussion of the results.
4.1. Poverty During Transition
Although the transition to a market economy has had severe effects on the living
conditions of the population through the lifting of price controls, the imposition of fiscal
discipline and the cutting of subsidies on the prices of basic commodities (Okrasa 1999b), it
has to be stressed that poverty, unlike open unemployment, is not a new phenomenon to
Polish society. During the 1970s and 1980s, income inequality was already higher in Poland
than in other middle and eastern European countries (see Golinowska 1998 and references
therein). Szulc (1997) argues that absolute poverty in Poland was higher in the 1980s than in
the 1990s despite the increase in official unemployment1. According to Keane and Prasad
(2002) as well as Sibley and Walsh (2002), the Gini estimate of income inequality hardly
changed over the 1990s at the national level.
However, Golinowska (1997) states that the 1990s are marked by an increase in
inequality and hence in relative poverty. In 1994 wages and real income rose again for the
first time since 1989 (Golinowska 1998). Nevertheless, mobility out of poverty was limited.
The fraction of those who remained poor for two consecutive years according to the official
92
1 Unemployment jumped from being unreported to about 17% in 1993 (Okrasa 1999b).

social-minimum criterion2 increased from 45% in 1988/89 to 72% in the recession period
1991/1992 (Okrasa 1999b). During the growth period of 1993 to 1996, the percentage
stabilized at about 60%3. That is, the pool of poor people seems to have been rather
stagnant. In industrial regions, it consists mainly of (former) blue collar workers, as Kotowska
(1997, 1998) shows. Despite the difficulties faced by industrial blue-collar workers, the most
dramatic decline in average incomes in the early transition phase was observed in farmers'
households (37% between 1988 and 1992) (Golinowska 1997). Moreover, households
associated with agriculture had a greater risk of falling into long-term poverty than employee
or pensioner households (Okrasa 1999b).
Poverty is also strongly related to unemployment. Of all households with at least one
unemployed member, about 27% experience poverty according to the relative poverty line,
whereas the share is 10% for families not affected by unemployment. The poverty rate is
even higher if unemployment benefits are the main source of income (Golinowska 1998).
Following Okrasa (1999b), long-term unemployment seems to be associated with chronic
poverty.
According to the World Bank, regional variation in poverty incidence, though existent is
not much pronounced. When analysing nine regions covering between four to eight
voivodships (Polish regions), the capital city region exhibits the lowest poverty incidence,
whereas in the Southeast and Central-West regions it is highest. Differences were a bit more
pronounced according to the human development index used by the United Nations.
Accordingly, sub-regional disparities in human development increased in Poland between
1992 and 1995, though the pattern remained unchanged (United Nations 1998). The lowest
level of human development was found in the Southeast and on the Baltic coast characterised
by many rural areas, the highest in large urban centres. In addition, while controlling for
regional fixed effects, Sibley and Walsh (2002) report measures of earnings inequality to be
higher in regions that are more advanced in restructuring.
The poverty status of a household is strongly affected by the level of education achieved
by the household head. For holders of a university diploma, the risk of falling into poverty is
three times as low as for other households (Okrasa 1999a). Using the household budget
survey data, the World Bank finds a strong inverse relationship between poverty and
education level of the household head. Post-secondary or university education even seem to
guarantee a living standard above the poverty line. At the lower end of the education scale,
however, the World Bank finds that the poverty rates of various social groups are quite
similar. The%age of people below the poverty line ranges between 23 and 27 when the
household head has only an elementary level of education no matter whether he is a worker
or a farmer.
93
4. Family Background and Children’s Educational ...
2 A low-income threshold had been introduced in 1981 as equal to the 1980 social minimum. It was updated thereafter.
The low-income threshold was quite similar to the relative poverty line of half-mean income.
3 Based on the relative poverty line of 50% of mean income in Poland for this period.

To summarise, the households who are most affected by poverty are those living in rural
areas, headed by poorly educated people and with unemployed members (see also Okrasa
1999a and 1999b). Moreover, households who prove to be less successful in avoiding
poverty are those with a high number of children. A World Bank study (Okrasa 1999b) states
that long-term poverty is a generation-skewed phenomenon in the sense that the number of
years in poverty grows with the number of children. Poland has an over-proportional share
of poor children. Golinowska (1997) also shows that the poverty rate increases as the
number of children in the family gets higher. In 1996, the poverty rate measured by the
minimum existence level criterion was 5.3% for families with 3 children, and 16% when the
number of children exceeded 4. Overall, the transition process has considerably increased
the social hardship of children (Golinowska 1996 and Unicef 1997).
The overall rise in disparities with respect to income as well as education across Polish
families, accompanied by higher costs of education for families, leads to the fear that in the
sequel children and young people will face increasingly unequal opportunities (Golinowska
1998). This problem is expected to be most severe for children with an agricultural
background. Not only does the typical poor Polish family live in an agricultural region and
have many children, but the children themselves also face worse educational prospects than
their peers in urban areas, partly due to poorer access to education institutions. The United
Nations (1998) report that 'young people of [rural], origin constitute half of the 19-24 age
group, but only 2% of tertiary students of that age group'.
Finally, there is evidence that the number of children is strongly related to the education
level of the parents. More than 60% of parents with three and more children have no
secondary education, whereas the overall average is 50% (Golinowska 1998). As mentioned
above, the poverty status of a household depends on the education level of the household
head. At the same time poverty is a question of the number of children and this is in turn
related to parents' education. It thus becomes evident that the education of parents and
children and other household characteristics reinforce each other, thereby creating a vicious
circle of economic deprivation.
4.2. Children's Education as a Mechanism of Poverty Transmission
The extent of poverty and the increase in unemployment arising with the transition
process might have far-reaching consequences for the Polish society. To the extent to which
poverty is transmitted from one generation to the next, the transition process will have
repercussions which go far beyond the transition period itself. Following the human capital
theory of Becker (1964) and Mincer (1974), education is a key element which affects
94

individual earnings prospects. In a broader perspective, educational attainment may be
viewed as the primary key to socio-economic success. Indeed, persons with a higher level of
education not only have higher earnings prospects when they work, but are also less likely
to experience unemployment and social exclusion in general. This link between educational
attainment and socio-economic outcomes has been confirmed by numerous empirical
studies for a wide range of countries (see for instance a review of the related literature for
Europe in Asplund and Pereira (1999)).
For Poland also, it has been found that poor educational achievement strongly enhances
the risk of experiencing poverty (World Bank 1994, 1995). Consequently, if parental poverty
– understood in a broad sense – has a large impact on the educational prospects of the
children, it enhances children's future poverty risk and poverty is likely to be passed on over
generations. Considering that the gains from education go far beyond the labour market and
have a wide range of positive externalities (on health, criminality, see among others Mayer
1994), it is of substantial interest to analyse the extent to which family background affects
one's educational prospects.
Family background may affect children's educational outcomes through various channels.
First, children growing in families where the education level of the parents is high might have
better educational prospects because they inherit to some extent the learning ability and
some other cultural endowments of their parents. Thus, Becker and Tomes (1986), and more
recently Ermisch and Francesconi (2001), point to the fact that part of a child's human capital
is "inherited" through the transmission of genetic and cultural endowments from parents to
children. The greater the degree of inheritability, the more closely related the human capital
of parents and children are. Sociologists and psychologists insist on the role of peer effects,
meaning that adults or peers to whom children relate set norms of desirable behaviour and
achievement (see Havemann and Wolfe 1995). Thus, the educational attainment of the
children is influenced by their social background through the transmission of ability as well as
of certain patterns of behaviour, preferences and expectations which, to some extent, are
internalised by children as standards and affect their cognitive and social-psychological
development.
Moreover, parental background might affect offspring's educational outcomes through
the availability of financial resources within the family. The acquisition of education may be
viewed as an investment, since it entails costs4 in the hope that it will bring about enhanced
earnings in the future (Becker 1964). For initial education, the investment has to be financed
by the parents, who are supposed to have an interest in the well-being of their offspring, i.e.
to be altruistic to some extent. Thus, children's educational outcomes are dependent on
intra-family transfers. In the presence of imperfect capital markets, and if the parents do not
95
4 The direct costs which might be associated with education, but also the opportunity costs caused by the time devoted
to education which is diverted from the labour market and potential earnings.

dispose of enough money – either because their wages are too low, or because they do not
dispose of any earnings nor other sources of income – investment in education might be
limited by credit constraints (see Rosenzweig and Wolpin 1993).
The positive correlation between family income and schooling attainment is well
documented in the literature, essentially for the United States (Solon 1992, Hill and Duncan
1987, Taubmann 1989) and has been widely interpreted as evidence of borrowing
constraints5. If parental income also proves to be correlated with children's education for
Poland, then the extent of poverty and unemployment, for instance, should have
repercussions on the educational attainment of the subsequent generations.
Not only the amount of family resources allocated to the children, but also the nature of
these resources and the timing of their distribution influence children's attainment. Thus,
children will also be affected by decisions such as the number of siblings, the region where
the family lives or the family structure (Haveman and Wolfe 1995). Because they generally
have fewer potential wage-earners, single-parent families have less access to financial
resources than two-parent families (Boggess 1998). They also have less time to spend on
their children, supervising their behaviour or assisting them in their school work. A working
mother also has less time to devote to her children, but in this case, there is a trade-off
between monetary and time resources. The increased parental income associated with the
mother's work might offset the reduction in child care time. In addition there is a cultural
aspect of the mother's job status that might have an impact on (particularly female) children's
education. Here, the evidence is contradictory, since Hill and Duncan (1987), for instance,
find a negative relationship between completed education and mother's work hours,
whereas Boggess (1998) does not find any significant impact of mother's employment on
children's educational attainment.
There is little evidence on the impact of family background variables like parental education,
occupational status or family structure for Poland. Among the few exceptions, Heyns and
Bialecki (1993) examine the impact of socio-economic background, as measured by father's
education and occupational prestige, on educational attainment for cohorts born between
1920 and 1969. The authors note, that upward social mobility is traditionally low in Poland and
children largely replicate the educational attainments of their parents, which is also argued by
the United Nations (1998). Heyns and Bialecki (1993) find that the effect of parental status
does not vary across cohorts, and that father's education is a far stronger variable in predicting
the educational attainment of Polish children than occupational variables. The authors conclude
that the increase in access to education seems to result from a more widespread availability of
schooling rather than from a change in the socio-economic determinants of educational
96
5 However, recently, some studies (e.g. Cameron and Heckman (1998), Shea (2000)) have contested the causal nature of
the link between family income and children's educational attainment, arguing that not parental income per se generates higher
educational achievement, but rather learning ability, in the sense that the commonly observed effect of parental income would
only reflect the correlation between parental income and parental ability, which, in turn, is correlated with children's ability.

success. This results contrasts with the analysis of the United Nations (1998), which, however,
covers a different time period, namely the transition period. Indeed, according to the report of
the United Nations (1998), even though the general level of education has increased in Poland
since 1989, with a particularly strong expansion of enrolments in higher education, there is
evidence that this educational expansion went along with an increase in social disparities in
access to education. The report also points to an increase in regional disparities, since young
people coming from rural areas have an increasingly difficult access to education at all levels.
4.3. An Empirical Analysis of Children's Educational Outcomes
4.3.1. The Polish Education System
In the following, the Polish education system is presented as it was organised throughout the
period under consideration (see Figure 4.1). The exposition largely follows that of the Education
Information Network in Europe (Eurydice 1999). In 1999, a reform of the education system has
taken effect. Structural reforms at the primary level have already been introduced in 1999/2000
whereas the reform of the upper secondary education level will start in the school year
2002/2003. Since the data available to us only cover the pre-reform period, we will restrict our
description to the old system. However, we will hint at changes where applicable.
97
Figure 4.1. The Polish education system
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9876543
Pre-primary education
Przedszkole
Basic vocat. school
Szko³a podstawowa
General & Vocat.
second. school
Technikum
General & Vocat. second.
school Liceum techniczne
zawodowe
Doctorates studies
General second. school
Liceum ogólnokszta³c¹ce
Higher vocational
schools
Post-secondary
Licentiate
degrees
University Master’s
degree
Pre-school
Primary education
Szko³a podstawowa
2nd stage
1st stage

The first level of the Polish school system is the mandatory pre-primary education for
children aged 3 to 6 in nursery schools and pre-school classes (oddzia³y przedszkolne)
attached to primary schools. In the latter six year old children have the right to complete a
year of preparation for primary education.
Primary education
Primary schools are divided into two stages: the first stage (grades 1 to 3) offering
elementary (block) learning and the second stage (grades 4 to 8) at which systematic
teaching is provided. Children leave primary school at age 14. There is no leaving
examination; children receive the primary school leaving certificate only (oewiadectwo
ukoñczenia szko³y podstawowej).
Secondary education
Secondary education covers the age group 15 to 18 or 19 (20). After the completion of
the 8-year single structure primary school, pupils have a choice between the following
schools:
– liceum ogólnokszta³c¹ce (4-year general secondary education for students aged 15 to 19)
– liceum zawodowe (4-year general and vocational secondary education for students aged
15 to 19)
– liceum techniczne (4-year secondary education supplemented with general and general
vocational subjects for students aged 15 to 19. This school type is comparatively new,
established only in the school year 1995/1996.)
– technikum (5-year vocational and technical secondary education for students aged 15 to 20)
– szko³a zasadnicza (3-year basic vocational education for students aged 15 to 18)
At the end of the first four types of schools6 pupils may take the matura examination
(egzamin dojrza³ooeci), which gives them the right for admission to higher education. Graduates
from upper secondary schools in Poland have a wide variety of educational possibilities at the
level of tertiary education. Those who do not pass the matura examination or who were not
accepted by higher education institutions, may continue their education in post-secondary
schools. Basic vocational graduates receive the qualification of skilled workers7.
Post-secondary education
Post-secondary schools (szko³a policealna), of 1 to 2.5 year duration, are considered as
part of secondary education in the Polish classification because of the type of qualifications
98
6 These types of educational institutions will start to be replaced by Liceum profilowane with the school year 2002/2003
and will completely disappear in 2004 respectively 2005.
7 With the school year 2002/2003 this institution will be changed to szko³a zawodowa in the reformed school system.

they offer. They prepare students for professional life. Students in these schools are trained
as nurses, accountants, administrative personnel for enterprises and hotels, computer
specialists or librarians. Those who have completed a course of study for a blue-collar
occupation obtain the title of skilled worker (robotnik wykwalifikowany) in the acquired
profession. Those who have completed a 2-year or 2.5-year course of study or a non-worker
specialization obtain the title of technician (technik) or an equivalent title.
Higher education
There are various types of non-university and university higher education institutions:
Teacher training colleges, traditional universities (uniwersytet), technical universities
(politechnika) and academies (akademia). At the end of 3 to 4 year higher vocational
education, students are awarded the vocational qualification diploma and the title of licencjat
or in¿ynier (both corresponding to bachelor, depending on the branch of study) which gives
them access to the job market or to extended higher studies. Universities and university-type
institutions with uniform master-degree studies of 4.5 to 6 year duration are entitled to
award the professional titles of magister (master), magister in¿ynier (master-engineer), lekarz
(doctor of medicine). Successful graduates can apply to do a doctorate.
Barriers in access to education / Public expenditures for education
Liberalisation and privatisation, as two systemic changes accompanying the
transformation to a market economy, have had diverse effects on the merits and costs of
education in Poland. While liberalisation of the labour market led to higher returns to
education in terms of job prospects and wages, thus increasing the incentive to invest in
education, privatisation of the education system resulted in a diversified quantity as well as
quality of educational institutions, particularly outside urban areas (United Nations 1998).
What has been diversified at the same time, however, is access to education which now
more and more depends on the income level of the parents due to the decline of state
expenditures and decentralisation of education services together with inadequate funding at
the local level and an increase of the costs to be covered by children's parents. In 1992 state
support for education was only 80% of that in 1989, in 1996 it amounted to 90% (United
Nations 1998).
4.3.2. Enrolments and Educational Attainment During Transition
In order to investigate the structure and the determinants of children's educational
attainment in Poland, we use data from LFS for the years 1992 to 2000. This way, we are able
to cover most of the transition process, though with less emphasis on the early recession period.
99

LFS is conducted as a national panel survey every three months (Szarkowski and
Witkowski 1994). After four preliminary quarters starting in May 1992, and repeated on ever
the same sample of households, a rotation system has been introduced in May 1993.
According to this system, in each quarter one completely new sample of housing units is
selected by two-stage sampling. Each sample is used following a 2-(2)-2 rule. This means that
a selected household stays in the survey for two quarters, is out for the next two quarters
and back again for another two quarters before it is finally discharged. The survey generally
covers all persons aged 15 and above. The respondents fill in two questionnaires: the first
one asks for general characteristics at the household level, registering all household
members (including children) and gathering information on the housing circumstances and
the family relations of all members. The second questionnaire covers only those persons
aged 15 and above living in the household. It collects information on socio-demographic and
labour-market characteristics at the individual level.
To draw a picture of the changes in the education of young people in Poland during the
1990s and to give a first illustration of the factors mentioned above that are potentially
related to it, we are now going to present some stylised facts on enrolments and educational
attainment.
Figure 4.2. Enrolment rate of 15-30 year-olds over time (1992-2000)
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Age
Figure 4.2 depicts the age-profile of participation in education in Poland8. The different
curves correspond to the years of the survey ranging from 1992 to 2000. As expected,
before the age of 15, almost all children are enrolled in education. Henceforth, the
enrolment rates are start declining. The decline is characterised by several non-linearities,
100
Source: Own calculations based on data from LFS, waves 1993-2000.
1992
1993
1994
1995
1996
1997
1998
1999
2000
8 The precise question in the PLFS we refer to is: "Are you a student of a day school or university?".

with one major drop in enrolments occurring above age 18 when basic vocational training is
completed. Beyond the age of 30, nearly all individuals have left the education system. From
the graphs, it is pretty obvious that enrolments have increased steadily over the past decade,
at least until 1999 and particularly for the age range 18 to 24. This means that on average
Polish people now study longer than they used to at the beginning of the decade, although,
in 2000, the enrolment of 17 to 21-year olds has decreased again, even below the 1998 level
for those of age 21. Under 18 and above age 24, however, the pattern of educational
participation has remained relatively stable over time. Lower secondary education has
always been on the agenda for teenagers. It seems that there has been a constant upward
shift in upper secondary and post-secondary education over the years, though. Whereas in
1992 only 84% of 18 year-olds and 20% of those aged 21 were studying, in 1999 the
respective numbers amounted to 95 and 47%. Higher education enrolments only started to
increase significantly in 1995, with yet another shift in 1999.
After having seen that participation in education has expanded during the 1990s and that
average schooling duration has become longer, we are now interested whether this
quantitative expansion of educational participation translated into a qualitative upgrade of
educational attainment in the course of the transition from a centralised to a market
economy. In addition, we want to get some first insights into the extent of intergenerational
mobility and examine the correlation between the highest education level attained by
individuals and some essential characteristics of their family background such as parents'
education and income as well as the region of residence.
The LFS does not ask questions on the respondents' parents. However, by means of a
household identification number, it is possible to match individuals with their parents
provided they live in the same household. Since the likelihood of having left the parental
household and living on one's own rises with age, we focus on younger individuals. For the
101
Education level variable Highest degree obtained Enrolled
Level 4: Higher education University
Post-secondary/vocational/
general school
yes or no
yes
Level 3: Secondary education Post-secondary school
Vocational school
General school
Basic vocational school
Elementary school
no
no
no
yes
yes
Level 2: Basic vocational education Basic vocational school no
Level 1: Primary education Elementary school
Less than elementary school
no
yes or no
Table 4.1. Construction of the education level variable

purpose of the analysis we have drawn a sample of individuals whom we could link to their
parents, young enough to minimise sample selectivity problems, but at the same time old
enough to have finished education or be about to finish it. After trying various ages for the
definition of the sample, we have finally selected a sample of 21 year-olds, for which a
reasonable proportion of which (about 65%) we are able to gather information on their
family background. Since a non-negligible part of the 21 year-olds has not yet finished
education at that age and is still enrolled in education, we need to take this into account. The
PFLS provides information on the highest degree obtained, and also whether the person is
currently enrolled in education or not, though not the specific level of education the person
is enrolled in. We combine this information to construct four levels of educational
attainment, ordered by level, as illustrated in Table 4.1.
First of all, we examine the structure of the highest degrees obtained by our sample of
21year-olds over the time period 1993 to 2000. Observations from the first wave gathered
in 1992 could not be considered since no question concerning the highest level of education
that the respondent had attained was asked in that year.
Figure 4.3. Highest education level of 21 year-olds over time (1993-2000)
27,4 28,7 27,2 31,1 21,9 21,8 22,9 24,4
33,0 29, 6 28,2 24,4
6,1 6, 6 8,5 8, 9
36,0 36,4 33,8 32,2
10,3 10,4 9,6 12,0
100%
90 %
80 %
70 %
60 %
50 %
40 %
30 %
20 %
10 %
From Figure 4.3, it appears that until 1997 the educational distribution is strongly
concentrated around intermediate qualification levels. Thus, the bulk of young Polish
people has received basic vocational education or upper secondary education at most,
while comparatively few people have attained higher education and even fewer hold a
primary education degree only. Considering developments over time, we can see that
there has been an upward shift during the period observed, with a decreasing percentage
of persons with poorer educational attainment and an increasing proportion of persons
102
35,6
33,5 35,1 36,2
33,8 31,4 31,9 31,4
0%
1993 1994 1995 1996 1997 1998 1999 2000
Year
Higher education
Upper secondary educ.
Basic vocational educ.
Primary educ ation

with a higher educational attainment. Looking more in detail, it seems that the
educational distribution has remained rather stable until 1995 and that an educational
upgrade has only occurred since then, presumably because of the expansion of private
education institutions. In particular, the proportion of graduates from tertiary level
institutions has increased strongly. In 2000, almost a third of the generation born in 1979
has reached the university level. At the same time, the proportion of basic vocational
certificate holders has decreased, especially since 1996. The proportion of 21 year-olds
with completed primary education only has also decreased between 1996 and 1997 but
is now slightly increasing again.
In the following figures, we examine the structure of educational attainment depending
on certain characteristics. Hereby, we pool the data over the period 1993 to 2000 and
observe the average educational distribution in this period for different groups of individuals.
Figure 4.4 shows the difference between 21 year old men and women in terms of
educational attainment.
Figure 4.4. Highest education level of 21 year-olds by gender (1993-2000)
19,3
27,8
39,8
27,1
37,1
26,9
13,1 8,9
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Obviously, 21 year-old women have reached a higher education level than their male
counterparts: The proportion of women with a higher education degree and with an upper
secondary education degree is by far higher than among men, while the percentage of
women having a basic vocational degree or having left school after primary education is
significantly lower than for men. These differences may partly be due to the types of
occupations young women and men choose, as most of the male-dominated jobs require
only basic vocational training whereas female-dominated occupations are typically preceded
by general secondary education.
103
Higher education
Upper secondary education
Basic vocational education
Primary education
Men Women

As we learned in Section 4.1, it seems that there exist some regional disparities in
educational achievement. Therefore, we look at the distribution of educational attainment
by region of residence in our sample of 21 year-olds over the time period 1993 to 2000. We
thereby distinguish 16 provinces that approximately equal the new Polish voivodships after
the territorial reform in 19999.
Figure 4.5. Highest education level of 21 year-olds by voivodship (1993-2000)
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Region
As appears from Figure 4.5, overall, the distribution of education looks quite similar
across regions, with the same concentration around intermediate qualification levels and
comparatively few people with a very high or a very low education level. However, some
provinces are outstanding. In the North-East, the proportion of tertiary level graduates is
particularly low (Warminsko-Mazurskie is the only voivodship where it is below 20%),
while the proportion of young people who completed primary education at most is
highest. The lowest proportion of 21-year olds with upper secondary or higher education
can be observed in Zachodniopomorskie (North-West). At the other end, the voivodship
Slaskie in the South has the highest percentage of graduates from higher education and the
lowest with only primary education, followed by Pomorskie in the North-West and the
capital region Mazowieckie. Hence, there is not much accordance between these
education levels and the picture drawn by the regional poverty rates of the World Bank
report. Only the capital city region seems to combine low poverty with a high proportion
of high level education.
104
Higher education
Primary education
Dolnooel¹skie
Kujawsko-Pomorskie
£ódzkie
Lubelskie
Lubuskie
Maopolskie
Mazowieskie
Podkarpackie
Podlaskie
Pomorskie
Opolskie
OEl¹skie
OEwiêtokrzyskie
Wielkopolskie
Zachodniopomorskie
9 New voivodship information is only given in the 2000 wave of the PLFS. For the preceding years we aggregated all old
49 voivodships to have a comparable measure. But as the new provinces are based on counties (powiaty) instead of old
voivodships we could not always ensure a 100% equivalence.

Figure 4.6. Highest education level of 21 year-olds by city size (1993-2000)
22,0 28,1 30,1
6,8 7,1 7,7 12,1
100%
90 %
80 %
70 %
60 %
50 %
40 %
30 %
20 %
10 %
More than region, the size of the city of residence seems to matter (see Figure 4.6). There
is a clear and strong relationship between the number of inhabitants of the city of residence
and the level of educational attainment: in large cities (with more than 100 inhabitants), the
proportion of tertiary level graduates is about 3 times as high as in rural areas, while the
percentage of 21-year olds who left school after primary education is only about half.
The proportion of children with a basic vocational degree at most is also particularly
high in rural areas, while the proportion of young persons with upper secondary education
is in general significantly lower than in urban areas, though about the same size as in big
cities and not to the same extent as that of higher education graduates. This picture is most
likely due to the supply of educational institutions dependent on city size. While
universities are located in larger cities, rural areas are not so well endowed, at least in
terms of public education institutions.
Since we are interested in the link between parental background and their offspring's
educational outcomes, we now examine the correlation between some essential parental
characteristics and the highest degree obtained by the children.
Figure 4.7 depicts the level of education of 21 year-olds depending on the education
level of their parents. Here, parents' education refers to the education level of that parent
holding the highest degree and is measured by four categories: primary, basic vocational,
secondary and higher education. As parents of 21 year-olds have generally finished their
education already, the differentiation is a bit different from that of their children and the
variable on the education level of the parents can easily be computed on the basis of the
highest school degree reported. As appears from Figure 4.7, there is a clear positive
105
41,2
30,7
37,3 37,2
32,9
40,6
27,6 25,0
13,8
0%
more than
100.000
20.000 -
100.000
less than
20.000
rural area
Population size
Higher education
Primary education

Figure 4.7. Highest education level of 21 year-olds by parents' education (1993-2000)
21,2
4,7
45,3
37,1
10 0%
90 %
80 %
70 %
60 %
50 %
40 %
30 %
20 %
10 %
correlation between parents' and children's education. While the numbers do not vary that
much for upper secondary education, the percentage of youngsters with higher education
but also primary and basic vocational education differ remarkably depending on the
parents' human capital endowment. For instance, more than two thirds of the individuals
with at least one parent having a higher education degree are themselves enrolled in
tertiary level studies, while this applies to less than 10% of the sons and daughters of
poorly educated parents (having completed at most primary education). Conversely,
almost two thirds of the young persons whose parents have only completed primary
education hold a basic vocational degree or less, while this concerns only 8% of those
persons with highly educated parents.
Furthermore, we look at the relationship between parental income and children's
educational attainment. Unfortunately, no information on total household income has been
collected in LFS. However, we have information on parents' labour income, but also on the
main source of income in the household. Figure 4.8 illustrates the relationship between
children's highest degree and total labour income in the household. Due to a currency
reform in Poland and due to high inflation rates, income is not easily comparable between
waves. We therefore use a relative measure of income indicating in which tertile (33,33%
quantile) of the earnings income distribution the household finds itself or whether the
parents receive no labour income at all.
As can be seen from Figure 4.8, having parents who fall into the lowest labour income
tertile is associated with a lower probability of achieving higher education and a
significantly higher probability of achieving only a basic vocational degree or less. Having
parents who are in the highest labour income tertile does improve educational prospects
106
5,6
27,5
37,4
37,8
24,2
9,8 17,8
36,2
67,7
17,4
0% 7,7 2,5
Higher education
Primary education
Primary
education
Basic
vocational
educ.
Upper
secendary
educ.
Higher
education
Parental education

Figure 4.8. Highest education level of 21 year-olds by household labour income
32, 5 34, 6
35, 2 37, 6
38, 5
24, 8 22, 4
11, 3 9, 4 5,2 4,9
100%
90 %
80 %
70 %
60 %
50 %
40 %
30 %
20 %
10 %
compared to the middle tertile, but only very slightly. Those persons living in households
with no labour income have slightly better educational prospects than those living in
households of the lowest labour income tertile. This might be explained by the fact that
part of the households without any labour income may receive income from other
sources, for instance from self-employment or state transfers. In order to consider this
issue, in Figure 4.9 we also examine the correlation between children's educational
prospects and the main source from which households draw their income, as this
information is also available in LFS.
107
34, 1
21, 0 18, 3
31, 4
38, 7
0%
Higher education
Primary education
No labour income Lowest tertile Middle tertille Highest tertille
Parental Labour income
Figure 4.9. Highest education level of 21 year-olds by main source of household income (1993-2000)
35.4
29.1
7.1
30.7
43.6
35.1
21.3
29.6
36.7
15.2 11.9
28.6
41.4 22.2
20.0 26.7
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
33.3
28.4
10.4
39.9
21.7 17.8
3.7
10.0
0%
Employment Farm
ownership
Self
employment
Pension Unemployment
benefits
Other
Higher education
Primary education

It appears that young persons living in households who draw their main income from
unemployment benefits or from farm ownership face the worst educational prospects, with
an extremely low probability of entering tertiary level studies and a particularly strong
probability of completing primary education or basic vocational education at most. On the
other hand, children in households having income from self-employment experience
significantly better educational prospects. Sons and daughters of pension beneficiaries seem
to have worse educational prospects than those of parents who dispose of labour income as
a main income source.
4.3.3. Determinants of Educational Attainment
While these simple correlations between the education of children and their respective
family background characteristics provide us with a first impression of the relations at work,
they may as well be misleading in the interpretation of social inequality in educational
attainment in Poland. Indeed, as conjectured above, background variables may be interrelated
and therefore reinforce each other. Given a negative correlation, say between the number of
children and their educational prospects for instance, one might wrongly attribute the
negative impact to the number of children while it actually stems from the fact that families
with many children generally live in the countryside and this is, in fact, the relevant factor
which negatively affects the educational prospects. Our analysis therefore aims at
disentangling the respective impacts of the various factors on educational attainment. In other
words, in our analysis of social inequality in educational attainment of young Polish people, we
now test whether the links brought up by the empirical literature and partly confirmed by our
figures still hold in a multivariate context. This will give us insights into true correlations as
opposed to spurious correlations. For this purpose, we set up an econometric model of
educational choice which accounts for the impact of various variables simultaneously.
In this model, we suppose that, for each individual, there exists an optimal amount of
education he or she would ideally like to attain, given some constraints, and let us call E* this
desired level of educational attainment. E* is a continuous variable which is not observable.
What can be observed is the actual decision of the individual given some characteristics, i.e.
the educational level E chosen among the J possible educational alternatives Ej which can be
ranked according to their levels, with j∈{1...J} and j=1 corresponding to lowest and j=J to
the highest educational level. The observable educational choice depends on the desired
level of schooling and on the opportunities available.
The decision on educational attainment is assumed to be rational in the sense that it
maximises the net perceived utility for the individual, subject to some constraints. Note that
it does not matter who in fact makes the decision, whether it is the individual himself or
108

i
somebody else ε(the parents, for instance). What counts is the outcome of the decision
among the possible alternatives. Let us suppose that, for each individual i, with i∈{1...N}, the
desired level of educational attainment can be expressed as a linear function of a vector of
individual characteristics xi and a residual term . Thus, we have:
Ei
*=βxi+εi
As mentioned previously, we do not observe the continuous variable Ei*, but the discrete
level Ei, which is defined to take a value Eij, with j∈{1...J}, if Ei* falls within a certain range
[μj-1, μj]:
Ei = EiJ if μJ Ei* > μJ-1
Ei(J-1) if μJ-1 Ei* > μJ-2
...
Ei2 if μ2 Ei* > μ1
Ei1 if μ1 Ei* > μ0
with μ1 = +∞ and μ0 = -∞.
Therefore, the probability that an individual i opts for educational level Eij given his/her
characteristics xi is:
Prob (Ei = Eij/x) = Prob (μj Ei* > μj-1)
= Prob ((μj βxi+εi > μj-1)
= Prob (μj -βxi εi > μj-1 -βxi)
Assuming that the residual terms εi
are normally distributed with mean 0 and variance σ2,
we obtain for all individuals i∈{1...N} and educational levels j∈{1...J}:
Prob (Ei =Eij/x) = Φ ( )− Φ ( )
where Φ is the cumulative standard normal distribution function.
This ordered probit model can only be identified up to a proportionality factor. Since we
can only identify the ratio of the parameters with respect to σ, it is usual in such models to
normalise σ to 1 (see Maddala 1983, p.23). The parameters β and the threshold values μ can
be estimated by maximising the likelihood function:
L =ΠΠ[Φ(μj -βxi) - Φ(μj-1 -βxi)]Iij
109
N J
i=1 j=1
μj-βxi σ
μj-1-βxi
σ

where Iij is an indicator variable equal to 1 if the individual i opts for educational level Ej
and 0 otherwise.
Maximising L boils down to maximising ln L since L is a positive function and ln L is a
monotone increasing transformation of L. Thus, the model can be estimated by maximising
the log-likelihood function:
N J
ln L=ΣΣ Iij ln[Φ(μj -βxi) - Φ(μj-1 -βxi)]
i=1 j=1
Table 4.2 presents the estimation results of this ordered probit model applied to the
sample of 21 year-olds from LFS for whom we could link the information of their parents.
The sample we have retained contains 4136 observations and covers the years between
1993 and 2000. The survey year 1992 could not be considered due to missing information
on the completed education level. The dependent variable is the highest education level
attained in four ordered levels as defined in Section 4.1. To stick to the model notation, E1 is
the lowest level of educational attainment and is defined as primary education or less, E2
corresponds to basic vocational education, E3 to upper secondary education and the highest
attainable education level E4 is defined as higher education.
As explanatory variables in the vector of characteristics xi, we use various indicators of
individual features: family structure, parents' human capital and parental wealth. As far as
individual characteristics are concerned, a dummy variable for sex is intended to assess the
nature and the extent of differences in educational opportunities for men and women.
Another dummy variable controls for the fact that disability might reduce educational
prospects. The number of children up to age 15 in the household is also included in the
regression as an indicator of family structure. This way, we hope to identify whether the
negative correlation often stated in empirical literature between the number of children in
the family and poverty (see Section 4.1) has for corollary a correlation between the number
of children and the educational attainment of children.
We add further information on city size, people living in rural areas building the reference
category in relation to which the coefficient estimates are to be interpreted. Fifteen region
dummies have also been included to check whether children living in certain parts of the
country are advantaged or disadvantaged compared to the reference category of those
residing in the OEl¹skie voivodship in the South of the country10,11. Parent's education is
represented by the highest educational level attained by the parents. By analogy with the
110
10 As in Section 4.2, the voivodship dummies have been built according to the new voivodship definition introduced in
1999. OEl¹skie has been chosen as the reference category because the percentage of our sample living there is the largest.
11 In sensitivity analyses we tried other regional differentiations, namely the nine macro regions used by the World Bank
and a classification of Polish regions into groups of different regional structure according to the methodology proposed by
Scarpetta and Huber (1995). Qualitatively, the results hardly vary between classifications.

descriptive overview in Section 4.1, we consider the education level of that parent holding
the highest degree.
As mentioned above, no information on total household income has been collected in LFS,
instead only labour income and the main source of income in the household are indicated.
Since, according to the literature cited above, this latter variable as well as parents' education
are strongly linked to poverty, these variables seem to be good proxies for household income.
Analogue to Section 4.1, we use a relative measure of parental labour income which indicates
the ratio of total parental labour income to the mean labour income of the year considered.
For an average income, the value of this variable will be one, for parental labour income above
(respectively below) average, it will be higher (respectively lower) than one. Since a rather
large proportion of our sample does not report any parental labour income, we also include
a dummy variable indicating this as an additional control variable. The labour income of the
parents might not be the main source of income of the household. Therefore, we also include
a set of dummy variables indicating where the main source of income of the household comes
from (farm ownership or farm use, self-employment, pension and unemployment benefits or
other non-earning sources), while income stemming from employment constitute the
references category (see also Figure 4.10).
In order to better assess the impact of parental unemployment on the educational
prospects of the children, we also include a dummy variable representing whether anyone of
the parents is currently unemployed. Better indicators of the family's financial background at
the time when the education decision of the child has been made would probably be the
incidence or duration of past unemployment spells of the parents. Unfortunately, this
information has not been gathered anymore in LFS in the years 1997 to 2000, so that we have
to draw on the incidence of current unemployment. Anyway, as a rule, current and past
unemployment are strongly correlated. Finally, time dummies have been included with a view
to capturing the extent of educational expansion over time.
The estimation results reveal that in the transition generation, women do better than men
in terms of educational attainment, other things equal. Being female is positively related to
one's education level at age 21, which is true at a significance level of 1%. The probability of
reaching a higher schooling level decreases significantly with the number of children and if the
person has a disability. Education is also positively related to city size as the education level of
21 year-olds is worst in rural areas, even if other characteristics are controlled for and is best
in cities with more than 100,000 inhabitants. The pattern of relative differences according to
city size remains similar to that observed on the bivariate level.
Remarkable disparities exist between regions. In Figure 4.5, it was not easy to compare
regions with each other: for instance, should a region which has a higher proportion of
university graduates than another, but at the same time more people with a very low
111

112
Table 4.2. Ordered probit estimation results
Dependent variable: Highest degree obtained.
Explanatory variables Coefficient Stand. dev.
Female (ref.: Male)
Disabled (ref.: Not disabled)
City size (ref.: Rural area)
<20,000 inhabitants
>100,000 inhabitants
Region (ref.: OEl¹skie)
Zachodniopomorskie
Pomorskie
Podlaskie
Lubuskie
Wielkopolskie
Kujawsko-pomorskie
Mazowieckie
Dolnooel¹skie
£ódzkie
Lubelskie
Opolskie
OEwiêtokrzyskie
Ma³opolskie
Podkarpackie
Highest education of parents (ref.: Primary education)
Secondary education
Higher education
Self-employment
Pension
Unemployment benefits
Parents have no labour income
Parents’ labour income (deviation from yearly mean)
At least one parent unemployed
Year of observation (ref.: 1993)
1994
1995
1996
1997
1998
1999
2000
0.500
-0.083
-1.202
0. 188
0. 216
0. 310
-0.385
-0.163
-0.187
-0.198
-0.326
-0.118
-0.117
-0.127
-0.253
-0.258
-0.019
-0.229
0.017
-0.139
-0.152
0.270
0.766
1.439
-0.092
0.236
-0.044
-0.152
0.132
0.124
-0.180
0.038
-0.038
0.044
0.205
0.207
0.196
0.267
***
***
***
***
***
***
***
*
*
*
***
*
***
***
**
*
***
***
***
***
*
*
***
***
***
***
***
0.040
0.023
0.108
0.058
0.050
0.049
0.093
0.088
0.098
0.114
0.100
0.075
0.083
0.069
0.079
0.080
0.086
0.108
0.092
0.079
0.095
0.046
0.049
0.084
0.069
0.073
0.054
0.116
0.065
0.050
0.063
0.072
0.080
0.071
0.070
0.070
0.070
0.074
Threshold values: μ1
μ2
μ3
-0.740
0.559
1.656
***
***
***
0.010
0.096
0.098
Log likelihood
Pseudo R²
Sample size
-4737.73
0.115
4136
Note: Level of statistical significance of the variables: * 10%, ** 5% and *** 1%.

education level be considered as better performing or not? The estimation results now
provide a kind of synthetic indicator which enables us to rank the regions in terms of
educational attainment. This ordinal ranking is illustrated through the shading of the Polish
voivodships in the map in Figure 4.10. The OEl¹skie voivodship chosen as the reference region
appears to be confirmed as the best-performing in terms of the educational attainment of its
youth, since all other regional dummies that are significant exhibit a negative sign. However,
five regions do not differ significantly from the OEl¹skie region (Wielkopolskie, Kujawsko-pomorskie,
Lubelskie, OEwiêtokrzyskie, Podkarpackie). The region which has the lowest
educational attainment of 21 year-olds seems to be the Zachodniopomorskie voivodship,
followed by £ódzkie, Dolnooel¹skie and Lubuskie.
Figure 4.10. Education level by voivodship, conditional on individual characteristics
The education of the parents seems to play a more important role than their employment
situation. Thus, children of parents with a higher education degree have by far the best
educational prospects, while those having parents having completed at most primary
education face the worst educational prospects. While the highest level of education
completed by parents is significantly positively linked to the educational attainment of their
offspring, there is only a weak, though significant, negative relation to parents' current
113
Note: The darker the shading of a region the lower the highest educational level of 21-year olds in that region, given the
socio-economic characteristics listed in Table 4.2.

unemployment. It has to be noted that the correlation observed on a bivariate basis between
unemployment and poor educational achievement is likely to be spurious for a part and may
essentially reflect the fact that parents experiencing unemployment are generally poorly
educated.
Parents' labour income seems to be significantly and positively correlated with the
educational achievement of the children, though the impact is limited in scope. The
relationship seemed to be stronger when only bivariate correlation was considered. This
means that part of this correlation actually stems from other variables, most probably from
the education of the parents which strongly determines labour income, rather than from the
income itself. The main source the household draws its income from does not seem to have
a significant influence on children's educational prospects when other factors are controlled
for, except that children of self-employed parents face significantly better educational
prospects than other children. Note that contrary to the simple correlation analysis above,
sons and daughters of households drawing their main income from farm ownership or farm
use are not significantly affected when other factors are controlled for (in particular city size
might be more determining, as well as parental education etc.), even though the effect might
be indirectly captured by the variable on parents' labour income, which has lower values for
agricultural households. The same is true for the offspring of pensioner households.
The year dummies show an effect only for the later transition period from 1997 onwards.
The positive signs of the years 1997 and following once again underline the pattern of
educational expansion in Poland, with particular jumps in 1997 and 2000, while controlling for
individual characteristics.
4.4. Conclusions
To sum up the results, the links between family background and children's education stated
in previous studies and the correlations found in the bivariate overview can only partly be
confirmed by our multivariate analysis. Similar to the results observed in the correlation
analysis, the children of self-employed seem to have better, those of unemployed worse
educational prospects, all else equal.
However, when controlling for personal characteristics such as sex, siblings, health status
and the parents' human capital endowment, parents' labour income is only weakly related to
the educational attainment of children. There may be different effects at work. First, the
observed bivariate correlation between parental income and children's education may in fact
be spurious and essentially the result of parents' education level, which is a strong determinant
of parental labour income. This could explain that when "cleaning" the income variable from
114

the education component, only a small effect remains for the income variable. An argument
for this is that the parental education variable still has a strong and very significant effect on
children's educational outcomes even though a variety of other characteristics are controlled
for. Nevertheless, the question remains whether the labour income variable represents a
reasonably good approximation of the household's wealth status. Also, the variable depicting
the main source of household income indicates the occupational position of the parents rather
than a true income effect and only imprecisely proxies the exact household income.
Moreover, children from farming families seem to be educationally disadvantaged, less so
for financial reasons than because of living in rural areas. The effect of city size is confirmed
even in a multivariate context as well as that of the region of residence. The Southern
provinces (OEl¹skie and adjacent voivodships to the East) of the country as well as
Wielkopolskie and Kujawsko-pomorskie remain those with the best educational prospects
whereas the region from Zachodniopomorskie to Dolnooel¹skie (in the North-West) and
£ódzkie cover those provinces with the worst educational prospects, but the ranking of the
other regions alters somewhat if other characteristics are controlled for. The bad educational
prospects of children living in the West as opposed to the East may be explained by the
provision of public schools in these voivodships. The results may be driven by our selection of
21-year olds who are still living with their parents. If, in the Western part of the country, young
people move from their parents' home earlier, for instance to take up studies in a voivodship
that provides a university or higher vocational school, the coefficient estimates are affected by
this selection. Finally, the negative impact of a large number of children often stated in the
literature remains remarkably significant also in the multivariate context, as well as the better
educational achievement of females compared to males.
Overall, in addition to revealing the links between children's education and their socio-economic
background, these results point to the usefulness of conducting multivariate
analyses instead of relying on bivariate correlations that can only tell us a part of the story.
115

Part 1
Allmendinger, J., S. Leibfried (2002), Education and Welfare State, [in:] European Review No. 10:3.
Atkinson, A. B. (1998), Poverty in Europe, Blackwell Publishers, Oxford.
Deniszczuk, L. (1978), Wzorzec konsumpcji spo³ecznie niezbêdnej [Pattern of Indispensable
Consumptions], Studia i Materia³y PIS, Zeszyt No 10, Warsaw.
Deniszczuk, L. (1998), Income, Spending and Savings in Households, [in:] S. Golinowska (ed.),
Social Policy in the 1990s, IPiSS, Warsaw.
Deniszczuk, L. (2002), Zró¿nicowanie dochodów i konsumpcji w Polsce [Income and
Consumption Differentiation in Poland], [in:] T. Kowalik (ed.), Nierówni i równiejsi,
Fundacja Innowacja, Warsaw.
ECLAC (2001), Social Panorama of Latin America, Santiago de Chile.
Erhel, Ch., J. Gautie, B. Gazier, S. Morel (1996), Job Opportunities for the Hard-to-Place, [in:]
G. Schmid, J. O'Reilly, K. Schoemann, International Handbook of Labour Market Policy and
Evaluation, EE, Cheltenham, Brookfield.
Fr¹ckiewicz, L. (1983), Sfery niedostatku [Spheres of Deficiency], Instytut Wydawniczy
Zwi¹zków Zawodowych, Warsaw.
Frieske, Kazimierz W. (ed.) (1999), Marginalnooeæ i procesy marginalizacji [Marginalisation and
Process of Marginalisation], Opracowania PBZ z. 13, IPiSS, IPS UW, Warsaw.
Golinowska, S. (ed.) (1996 and 1997), Polska bieda. Kryteria. Oceny. Przeciwdzia³anie [Polish Poverty.
Criteria. Evaluation. Counteracting Measures], Studia i Materia³y IPiSS, IPiSS, Warsaw.
Golinowska, S., I. Topiñska (2002), Pomoc spo³eczna – zmiany i warunki skutecznego dzia³ania
[Social Assistance – Changes and Conditions of the Effective Actions), CASE, Warsaw
GUS (1997), WskaŸniki ubóstwa w oewietle badañ bud¿etów gospodarstw domowych [Poverty
Indicators in View of Household Budget Surveys], GUS Warsaw.
117
References

GUS (1999), Warunki ¿ycia ludnooeci w 1999 r. [Living Conditions of the Populations in 1998],
Warsaw.
GUS (2000), Warunki ¿ycia ludnooeci w 1999 r. [Living Conditions of the Populations in 1999],
Warsaw.
Innocenti Research Centre of Unicef (1995), Poverty, Children and Policy: Responses for a
Brighter Future, Regional Monitoring Report No. 3, Florence.
Innocenti Research Centre of Unicef (2001), A Decade of Transition, The Monee Project
CEE/CIS/ Baltics, Regional Monitoring Report No. 8, Florence.
Jarosz, M. (1984), Nierównooeci spo³eczne [Social Inequalities], KIW, Warsaw.
Klugman, J., J. Micklewright, G. Redmond (2002), Poverty in the Transition: Social Expenditures
and the Working-Age-Poor, Innocenti Working Papers No. 91, Innocenti Research Centre
of Unicef, Florence.
Kordos, J. (1992), Poverty Measurement in Poland, [in:] Poverty Measurement for Economic in
Transition in Eastern European Countries, PTS/GUS, Warsaw.
Kraemer, W. (2000), Armut in der Bundesrepublik. Zur Theorie und Praxis eines ueberforderten
Begriff, Campus Verlag, Grankfurt-Main.
Kryñska, E., J. Suchecka, B. Suchecki (1999), Prognoza poda¿y i popytu na pracê do roku 2010
[Forecast of Labour Supply and Labour Demand Until 2010], IPiSS, Warsaw.
Leisering, L., S. Leibfried (1999), Time and Poverty in Western Welfare States. United Germany
Perspective, Cambridge University Press, Cambridge.
Leoe, E., S. Na³êcz (2001), Sektor non-profit. Nowe dane i nowe spojrzenie na spo³eczeñstwo
obywatelskie w Polsce [Non-Profit Sector. New Data and New Look at the Civil Society in
Poland], Instytut Studiów Politycznych PAN i John Hopkins University, Warsaw.
Lewis, O. (1959), Five Families. Mexican Case Studies in the Culture of Poverty, Science
Foundation, New York.
Lewis, O. (1966), Nagie ¿ycie [La Vida], PIW, Warsaw.
Lipton, M. (1995), Growing Points in Poverty Research: Labour Issues, [in:] G. Rodgers (ed.) New
Approches to Poverty. Analysis and Policy, I. The Poverty Agenda and the ILO, Issues for
Research and Action, ILO, Geneva.
McAuley, A. (1979), Economic Welfare in the Soviet Union: Poverty, Living Standard and
Inequality, University of Wisconsin Press, Madison, Wisconsin.
Milanovic, B. (1998), Income, Inequality and Poverty During the Transition from Planned to
Market Economy, The World Bank, Washington D.C.
Mingione, E. (1996), Urban Poverty in Advanced Industrial World: Concepts, Analysis and Debates,
[in:] E. Mingione (ed.), Urban Poverty and the Underclass, Basil Blackwell, Oxford.
Podgórski, J., B. Dobrowolska (1991), Subiektywne granice ubóstwa [Subjective Borders of
Poverty], GUS, Warsaw.
Podgórski, J. (1994), Wyznaczanie subiektywnych linii ubóstwa [Defining of the Subjective
118
CASE Reports No. 54 Poverty Dynamics in Poland

Poverty Lines], [in:] Wiadomooeci Statystyczne No. 12, p. 12-19.
Randtel, U., G. Wagner (eds) (1991), Lebenslagen im Wandel: Zur Einkommensdynamik in
Deutschland seit 1984, Campus Verlag, Frankfurt am Main.
Rowntree, B.S. (1901), Poverty. A study of Town Life, 2nd edition, Thomas Nelson & Sons,
London.
Tarkowska, E. (1998), Ubóstwo w by³ych PGR-ach. W poszukiwaniu dawnych Ÿróde³ nowej biedy
[Poverty in the Former State Farms – PGR. Looking for Old Causes of the New Poverty],
[in:] Kultura i Spo³eczeñstwo 1998 nr 2.
Tarkowska, E. (2000), Popegerowska bieda [Poverty after Collapse of Big State Farms – PGR],
[in:] E. Tarkowska (ed.), Zrozumieæ biednego. O dawnej i nowej biedzie w Polsce
[Understanding of the Poor Man. About the Old and the New Poverty), Typografika,
Warsaw.
Tarkowska, E., K. Korzeniewska (2001), M³odzi z dawnych PGR-ów [Youth from the Old State
Farms – PGR], Instytut Spraw Publicznych, Warsaw (forthcoming).
Topiñska, I. (1997), Ubóstwo w Polsce: 1993-1996 [Poverty in Poland: 1993-1996], [in:]
S. Golinowska (ed.), Polska bieda. Kryteria. Oceny. Przeciwdzialanie [Polish Poverty.
Criteria. Evaluation. Counteracting Measures], Studia i Materia³y IPiSS, IPiSS, Warsaw.
Tymowski, A. (1973), Minimum socjalne. Metodyka i próba okreoelenia [Social Minimum.
Methodology and Estimation Effort], PWE, Warsaw.
UNDP (2000), Rural Development. National Development Report Poland 2000, Warsaw.
Witkowski, J. (2001), Transformacja rynku pracy [Labour Market Transformation], Kontrola
Pañstwowa nr 3, Najwy¿sza Izba Kontroli, Warsaw.
World Bank (1990), Poverty World Development, Report 1990, Oxford University Press.
World Bank (1994), Poverty in Poland, Washington D.C.
World Bank (1995), Understanding Poverty in Poland, Washington D.C.
World Bank (2000), Making Transition Work for Everyone. Poverty and Inequality in Europe and
Central Asia, Washington D.C.
World Bank (2001), Poland: Labour Market Study. The Challenge of Job Creation and Distruction,
Washington D.C.
Part 2
Aksman, E. (2002), Redystrybucyjny efekt podatku dochodowego w Polsce [Redistributive Effects
of Personal Income Tax in Poland], Economista nr 4, Wydawnictwo Key Text, Warsaw.
Atkinson, A.B. (1992), Economic Transformation in Eastern Europe and the Distribution of
Income, Cambridge University.
Broda-Wysocki, P. (1999), Pomoc spo³eczna. Pomagaæ i motywowaæ do indywidualnej aktywnooeci
[Social Assistance. Help and Motivate to Individual Activity), [in:] S. Golinowska (ed.),
Zmiany i reformy w systemie zabezpieczenia spo³ecznego. W kierunku wzrostu indywidualnej
119
References

odpowiedzialnooeci [Reforms of the Social Security System. Towards Higher Individual
Responsibility], PBZ z. 16, IPiSS, Warsaw, p. 134-171.
Fortin, B., G. Lacroix (2002), Assessing the Impact of Tax and Transfers Policies on Labour
Supply: a Survey, CIRANO – Centre interuniversitaire de recherche en analyse des
organisations, Montreal.
Golinowska, S. (1996), Polityka spo³eczna w latach 1994-1996. Procesy regulacyjne i ich
potencjalne skutki [Social Policy in 1994-1996. Structural Changes and Potential
Consequencies], Raport IPiSS, Zeszyt nr 11, Warsaw.
Golinowska, S. (1998), Public Social Expenditures in Poland in the Period of Transition, Industrial
and Social Policies in Countries in Transition, University of Postdam, Postdam.
Golinowska, S. (1999), Zasi³ki dla bezrobotnych. Motywacja do wyjoecia z bezrobocia czy pozostania
w nim? [Unemployment Benefits. Motivation to Quit or Stay in Unemployment?], [in:]
Golinowska S. (ed.) Zmiany i reformy w systemie zabezpieczenia spo³ecznego. W kierunku
wzrostu indywidualnej odpowiedzialnooeci [Reforms of the Social Security System. Towards
Higher Individual Responsibility], z. 11, IPiSS, Warsaw, p. 107-131.
Golinowska, S. (2000a), Polityka spo³eczna; koncepcje-instytucje-koszty [Social Policy; Ideas-
Institutions-Costs], Poltext, Warsaw.
Golinowska S. (2000b), Public Social Expenditure in Poland in the Period of Transition, [in:]
H.-G. Petersen (ed.), Industrial and Social Policy in Transition Countries, Shaker Verlag
Aachen.
Golinowska, S. (2002), O wydatkach spo³ecznych w oewietle ustawy bud¿etowej na rok 2002 [About
the Social Expenditures in the Light of the Budget Law for 2002], [in:] Polska Gospodarka.
Tendencje. Oceny. Prognozy I/2002 (12) (Polish Economic Outlook), CASE, Warsaw.
Golinowska, S., K. Hagemejer (1999), Spo³eczne wydatki w Polsce z odniesieniami do innych
krajów [Social Expenditures in Poland in Comparison with Other Countries], z. 11, PBZ
Warsaw.
Górecki, B., M. Wioeniewski (1998), Zró¿nicowanie dochodów gospodarstw domowych w Polsce
w latach 1987-1995 [Differentiation of Household Incomes in 1987-1995], [in:] Podzia³
dochodu i nierównooeci dochodowe. Fakty, tendencje, porównania. [Income Distribution and
Income Disparities. Facts, Tendencies, Comparisons], Rada Strategii Spo³eczno-
Gospodarczej przy Radzie Ministrów, Warsaw, p. 133-170.
Gujarati, D.N. (1995), Basic econometrics, McGraw-Hill, Inc., Singapore.
GUS (1997 and 2001), Statistical Yearbook, GUS, Warsaw.
Koptas, G. (1995), Labour Market Policy and Help for the Unemployed, [in:] M. Bednarski (ed.),
Social policy in 1993-1994. Main problems in transition period, IPiSS, Warsaw.
Kramer, G., J. Neneman (1995), The Effects of Reforms of Income Taxation in Poland, [in:]
M. Belka, H-G. Petersen (eds.), Economic Transformation in Poland. Reforms of Institutional
Settings and Macroeconomic Performance, Campus Verlag, Frankfurt / New York, p. 239-261.
120

Kurzynowski, A. (1994), Polityka spo³eczna i jej przemiany w latach 1990-1994 – ogólny pogl¹d
[Social Policy in 1990-1994 – General Opinion], [in:] U. Grzeloñska, Polityka spo³eczna w
Polsce w okresie transformacji [Social Policy in Poland During Transition], Szko³a G³ówna
Handlowa, Warsaw.
MPiPS (2002), Transfery socjalne [Social Transfers], May 2002, Warsaw.
Nagel, T., J. Neneman (1995), The Effects of a Reform of Private Consumption Taxation in
Poland, [in:] M. Belka, H-G. Petersen (eds.), Economic Transformation in Poland. Reforms of
Institutional Settings and Macroeconomic Performance, Campus Verlag, Frankfurt / New
York, p. 205-237.
Perivier, H. (2001), Transferts redistributifs et offre de travail des ménages : analyse théorique et
mise en perspective franco-américaine [Redistributive Transfers and Labor Supply: a
Theoretical Analysis and a French-American Comparison], doctoral dissertation,
Sorbonne University.
Puhani, P. (1998), Advantage through training? A microeconometric evaluation of the employment
effects of active labour market programmes in Poland, Discussion Paper No. 98-25, Centre
for European Economic Research (ZEW), Mannheim.
Rutkowski, J.J. (1998), Welfare and the Labor Market in Poland. Social Policy During Economic
Transition, The World Bank, Washington D.C.
Styczeñ, M., I. Topiñska, (1999), Podatki i wydatki socjalne jako narzêdzia redystrybucji
dochodów gospodarstw domowych [Taxes and Social Expenditures as Instruments of
Redistribution], IPiSS, Warsaw.
Szylko-Skoczny, M. (ed.) (1993), Pomoc spo³eczna wobec zjawiska bezrobocia. Raport z badañ
[Social Assistance and Unemployment. Research Report], Friedrich Ebert Stiftung, Warsaw.
Topiñska, I. (1991), The Impact of Social Benefits and Subsidies on Income Distribution: Poland,
1990, report for MPiPS, Warsaw, mimeo.
ZUS (2001), Wa¿niejsze informacje z zakresu ubezpieczeñ spo³ecznych (FUS) 2000 r. [Basic
Information About the Employee Social Security System in 2000], ZUS, June, Warsaw.
ZUS (2001), Biuletyn informacyjny Nr 3 (6)/2001m [Information Bulletin], ZUS, Warsaw.
Part 3
Golinowska, S. (ed.) (1996), Polska bieda. Kryteria. Oceny. Przeciwdzia³anie [Polish Poverty.
Criteria. Evaluation. Counteracting Measures], IPiSS, Warsaw.
Golinowska, S., I. Topiñska (2002), Pomoc spo³eczna – zmiany i warunki skutecznego dzia³ania
[Social assistance – Changes and Conditions of the Effective Actions], CASE, Warsaw.
Groenmeyer, A. (1999), Alkohol, Alkoholkonsum und Alkoholprobleme, [in:] Handbuch Soziale
Probleme, Westdeutschen Verlag, Opladen.
GUS (2000), Wybrane aspekty stanu zdrowia ludnooeci Polski na tle krajów europejskich
[Selective Aspects of Health Condition of Polish Society Compared with European
121
References

Countries], Warsaw.
Klingemann, H. (2001), Public order and safety, [in:] Klingeman H., Gmel G. (eds.), Mapping
the Social Consequences of Alcohol Consumption, WHO regional Office for Europe, Kluwe
Academic Publishers.
Maffli, E. (2001), Problem drinking and relatives, [in:] Klingeman H., Gmel G. (eds), Mapping
the Social Consequences of Alcohol Consumption, WHO regional Office for Europe, Kluwe
Academic Publishers.
Mellibruda, J. (1999), Psychologiczna koncepcja mechanizmów uzale¿nienia [Psychological
Concept of Dependency Mechanisms], [in:] Terapia Uzale¿nienia i Wspó³uzale¿nienia
[Therapy of Dependence and Interdependence], No. 1(4).
Morawski, J., Moskalewicz J., Wald I. (1991), Economic Costs of Alcohol Abuse, with Special
Emphasis on Productivity, conference materials, Oslo.
Moskalewicz, J., J. Sieros³awski, T. Szymanowski, G. OEwi¹tkiewicz, A. Zieliñski (1991),
Alkohol jako problem spo³eczny [Alcohol as a Social Problem], [in:] Patologia spo³eczna.
Wybrane problemy [Social Pathologies. Selective Problems], WSPS, Warsaw.
PARPA (2001a), Polskie problemy alkoholowe [Polish Alcohol Problems], www.parpa.com.pl
PARPA (2001b), Statystyki i sprawozdania [Statistics and Reports], www.parpa.com.pl
Pernanen, K. (2001), Consequences of Drinking to Friends and the Close Social Environment,
[in:] Klingeman H., Gmel G. (eds.), Mapping the Social Consequences of Alcohol
Consumption, WHO regional Office for Europe, Kluwe Academic Publishers.
Rehm, J. (2001), Concepts, Dimensions and Measures of Alcohol-Related Social Consequences –
A basic Framework for Alcohol-Related Benefits and Harm, [in:] Klingeman H., Gmel G.
(eds.), Mapping the Social Consequences of Alcohol Consumption, WHO regional Office for
Europe, Kluwe Academic Publishers.
Rehm, J., I. Rossow (2001), The Impact of Alcohol Consumption on Work and Education, [in:]
Klingeman H., Gmel G. (eds.), Mapping the Social Consequences of Alcohol Consumption,
WHO regional Office for Europe, Kluwe Academic Publishers.
Schäfer, J. (1996), Alkohol und Arbeitsverhältnis. Zur rechtlichen Bewältigung alltäglicher
Probleme, Europäische Hochschulschriften, Frankfurt am Main.
Sieros³awski, J. (1998), Problemy alkoholowe w ocenie mieszkañców. Raport z ogólnopolskich
badañ ankietowych zrealizowanych w 1998 r. [Alcohol Problems in the Opinion of the
Society. A Report on the Survey Conducted in 1998], www.parpa.pl
Sieros³awski, J., T. Szumlicz (1995), Alkohol a problemy rodziny [Alcohol and Family Problems],
[in:] Rodziny w Polsce [Families in Poland], IPiSS, Warsaw.
WHO (2001), Global Status Report on Alcohol, www.who.int
WHO (2002), The World Health Report 2001. Mental Health: New Understanding, New Hope,
www.who.int
World Bank (1995), Understanding Poverty in Poland, Washington.
122

Part 4
Asplund, R., P.T. Pereira (1999), Returns to Human Capital in Europe, A Literature Review, B
Series, Vol. 156, Helsinki, p. 125-146.
Becker, G. (1964), Human Capital: A Theoretical and Empirical Analysis, With Special Reference
to Education, National Bureau of Economic Research, New York.
Becker, G.S., N. Tomes (1986), Human Capital and the Rise and Fall of Families, [in:] Journal of
Labor Economics, Vol. 4, p. 1-39.
Boggess, S. (1998), Family Structure, Economic Status, and Educational Attainment, [in:] Journal
of Population Economics, Vol. 11(2), 205-222.
Cameron, S.V., J. J. Heckman (1998), Life Cycle Schooling and Dynamic Selection Bias: Models
and Evidence for Five Cohorts of American Males, [in:] Journal of Political Economy, Vol. 106
(2), p. 262-333.
Ermisch, J.F., M. Francesconi (2001), Family Matters: Impacts of Family Background on
Educational Attainments, [in:] Economica, Vol. 68, p. 137-156.
Eurydice, (1999), Structures of Education, Initial Training and Adult Education Systems in
Europe, Eurydice, Warsaw,1-35.
Golinowska, S. (1996), Social Policy Towards Poverty. Comparative Approach, Friedrich-Ebert-
Stiftung, IPiSS, Warsaw.
Golinowska, S. (1997), Polska Bieda. Kryteria. Ocena. Przeciwdzia³anie [Polish Poverty.
Criteria. Evaluation. Counteracting Measures], IPiSS, Warsaw.
Golinowska, S. (1998), Armut in Polen, [in:] Voges W. (ed.), Armut in Europa, p. 196-215.
Haveman, R., B. Wolfe (1995), The Determinants of Children's Attainments: A Review of
Methods and Findings, [in:] The Journal of Economic Literature, Vol. 33 (4), 1829-1878.
Heyns, B., I. Bia³ecki (1993), Educational Inequalities in Postwar Poland, [in:] Y. Shavit and H.-
P. Blossfeld (eds.), Persistent Inequality, San Francisco, p. 303-335.
Hill, M., G. Duncan (1987), Parental Family Income and the Socio-Economic Attainment of
Children, [in:] Social Sciences Resources, Vol. 16, p. 37-73.
Keane, M., E. Prasad (2002), Inequality, Transfers and Growth: New Evidence from the Economic
Transition in Poland, IZA Discussion Paper No. 448, Bonn.
Kotowska, I.E. (1997), Reducing Income by Income Support in Poland. Some Results of the
Household Panel Based Study, Warsaw School of Economics, mimeo.
Kotowska, I.E. (1998), Living Conditions of Blue Collar Households – A Threat of Poverty,
Warsaw School of Economics, mimeo.
Maddala, G.S. (1983), Limited-Dependent and Qualitative Variables in Econometrics, Cambridge.
Mayer, H. (1994), Bildungsökonomie: die Interdependenz von Bildungs- und Beschäftigungssystem,
Stuttgart.
Mincer, J. (1974), Schooling, Experience and Earnings, National Bureau of Economic Research,
New York.
123
References

Okrasa, W. (1999a), The Dynamics of Poverty and the Effectiveness of Poland's Safety Net (1993-
1996), Policy Research Working Paper 2221, World Bank, Development Research Group,
Poverty and Human Resources, Washington D.C.
Okrasa, W. (1999b), Who Avoids and Who Escapes From Poverty During the Transition: Evidence
from Polish Panel Data 1993-1996, Policy Research working paper 2218, World Bank,
Development Research Group, Poverty and Human Resources, Washington D.C.
Rosenzweig, M.R., K.I. Wolpin (1993), Intergenerational Support and the Life-Cycle Incomes of
Young Men and Their Parents: Human Capital Investments, Coresidence and Intergenerational
Financial Transfers, [in:] Journal of Labor Economics, Vol. 11(1), p. 84-112.
Scarpetta, S., P. Huber (1995), Regional Economic Structures and Unemployment in Central and
Eastern Europe: An Attempt to Identify Common Patterns, [in:] The Regional Dimension of
Unemployment in Transition Countries, OECD.
Shea, J. (2000), Does Parents' Money Matter?, [in:] Journal of Public Economics, Vol. 77 (2), p. 55-184.
Sibley, C.W., P.P. Walsh (2002), Earnings Inequality and Transition: A Regional Analysis of Poland,
IZA Discussion Paper No. 441, Bonn.
Solon, G. (1992), Intergenerational Income Mobility in the United States, [in:] American
Economic Review, Vol. 82, p. 393-408.
Szarkowski, A., J. Witkowski (1994), The Polish Labour Force Survey, [in:] Statistics in
Transition, Vol. 1 (4), p. 467-483.
Szulc, A. (1997), Poverty and Equality in Poland in the 1990s, Warsaw School of Economics,
mimeo.
Taubmann, P. (1989), Role of Parental Income in Educational Attainment, [in:] AEA Papers and
Proceedings, Vol. 79(2), p. 57-61.
Unicef (1997), Children at Risk in Central and Eastern Europe: Perils and Promises, A Summary,
Florence.
United Nations (1998), Human Development Report Poland 1998.
World Bank (1994), Poverty in Poland. Report No.1, Washington D.C.
World Bank (1995), Understanding Poverty in Poland, Washington D.C.
124

CASE Network Report 54 - Poverty Dynamics in Poland. Selected Quantitative Analyses

More Related Content

What's hot (20)

Viewers also liked (13)

Similar to CASE Network Report 54 - Poverty Dynamics in Poland. Selected Quantitative Analyses (20)

More from CASE Center for Social and Economic Research (20)

Recently uploaded (20)

CASE Network Report 54 - Poverty Dynamics in Poland. Selected Quantitative Analyses