Day 2 (Room C, July 2) - Congress Replay - The 6th Congress of Polish Statistics

Objective

Characteristics of the measurement of economic poverty in Poland, taking into account the following issues: 1. Benchmarks for the objective measures of economic poverty used 2. Interpretation of the main indicators of economic poverty and privation 3. Data sources enabling the estimation of economic poverty and privation in Poland. 4. Trends in various types of economic poverty, privation and subjective assessments of financial circumstances 5. What scales are used in responses to questions regarding the subjective assessment of financial circumstances 6. Poverty and the social Cohesion

Methods

An analytical and synthetic approach to publications on research conducted by the Statistics Poland, from which the data used to estimate economic poverty are derived, as well as participant observation in the research process on poverty and social exclusion carried out by the Statistics Poland. The data sources on which the estimation of economic poverty is based are the results of the following representative surveys carried out by Statistics Poland: the annual Household Budget Survey (sample – approx. 30–35,000 households), the annual European Survey on the Living Conditions of the Population (approx. 14,000 households) and the periodic (every few years, with the latest edition of the survey in 2018) Social Cohesion Survey (current sample size – approx. 25–26,000 households). The research tools used in these surveys are questionnaires. The expenditure-based approach predominates in estimating the extent of the risk of economic poverty (this applies to the risk of extreme poverty, relative expenditure-based and legal poverty (which are estimated based on monthly expenditure on goods and services, including the renovation fund, using data from the Household Budget Survey) over the income-based approach (which is estimated based on annual income using data from EU-SILC and the Social Cohesion Survey).

Results

1. The ranges of risk associated with various types of economic poverty estimated in Poland, and the trends in these ranges over time, differ due to varying interpretations of the meaning of these types 2. The trends in extreme poverty and privation coincide (the constructs of both indicators are similar) 3. Trends in the prevalence of extreme poverty and assessments of a poor and very poor financial situation may show opposite dynamics 4. In subjective assessments for national purposes, a Likert scale is used instead of, for example, the 10-point scales used by Eurostat 5. Poverty was one of the elements of the analysis in the Social Cohesion Survey, which stemmed from the understanding of social cohesion adopted by Polish Statistics. UNECE ignores aspect of poverty in social cohesion.

Conclusions

1. In its study of poverty as a multidimensional phenomenon, Statistics Poland uses a relatively wide range of objective measures to estimate economic poverty and, as a supplement, also employs subjective assessments. 2. Each measure has a different interpretation and, from a methodological point of view, there is no basis for favouring any one of them. 3. In addition to the measures used to assess economic poverty, it would be worthwhile to introduce a measure that takes into account assets that can be quickly liquidated. The use of an indicator covering such assets would provide a more comprehensive picture of poverty. Legal poverty thresholds should be set more frequently than 3 years.

Keywords

the risk of extreme, relative and legal poverty: the risk of privation: the expenditure-based and income-based approaches:

Objective

The aim of this paper is to present the results of the estimation of the at-risk-of-poverty rate (AROP) in Poland at the subregion level (NUTS 3) for 2019–2023, using univariate (UFH) and multivariate (MFH) Fay-Herriot models. The study was carried out in cooperation between the Poznań University of Economics and Business, Statistics Poland, the Statistical Office in Poznań, and the World Bank. Small area estimation methods are applied to provide precise poverty estimates at a lower level of spatial aggregation than that previously published by official statistics.

Methods

The study covered all 73 subregions (NUTS 3) in Poland for the years 2019–2023. Direct estimates of AROP were obtained on the basis of the sample from the EU-SILC survey conducted by Statistics Poland, whose sample size and sampling design do not allow for the publication of reliable estimates below the regional level (provinces). In the Fay-Herriot models, the direct estimates of AROP were used as the dependent variable, while the auxiliary variables were taken from the Local Data Bank of Statistics Poland (BDL GUS) and described the demographic structure, the situation on the labour market, migration, income, and housing infrastructure (sets of 6 and 9 auxiliary variables were considered). In the UFH model, the empirical best linear unbiased predictor (EBLUP) was applied, which is a weighted combination of the direct and synthetic estimators. The multivariate Fay-Herriot (MFH) model extends the univariate approach by simultaneously taking into account a vector of correlated area-specific characteristics, which makes it possible to exploit the correlations between AROP and the auxiliary variables in order to improve the precision of the estimates. The model parameters were estimated by means of restricted maximum likelihood (REML), and the quality of the obtained estimates was assessed using the coefficient of variation (CV) and compared with the precision of the direct estimates.

Results

The poverty maps reveal a pronounced spatial variation of AROP in Poland: higher values are observed in central and eastern Poland, lower values in western Poland, and the lowest values in large cities and the subregions surrounding them. The estimates obtained on the basis of the UFH and MFH models are characterised by greater precision than the direct estimates: the coefficients of variation (CV) are noticeably lower throughout the entire analysed period of 2019-2023. The UFH models based on 6 and 9 auxiliary variables lead to similar results, while the application of MFH additionally improves the quality of the estimates by exploiting the correlations between area-specific characteristics, providing a coherent picture of the spatial differentiation of poverty in Poland.

Conclusions

The application of Fay-Herriot models makes it possible to estimate the AROP indicator at the level of 73 subregions (NUTS 3), for which results have not previously been published by official statistics in Poland. MFH constitutes an extension of the UFH approach and allows for the effective use of information derived from multiple correlated area-specific characteristics. The results of the study, carried out in cooperation with the World Bank, may be applied in the design of cohesion policy and the allocation of development funds. The proposed approach is transferable to other areas of official statistics in which there is a need for estimation at low levels of spatial aggregation.

Keywords

small area estimation, Fay-Herriot model, multivariate Fay-Herriot model, at-risk-of-poverty rate (AROP), poverty maps

Objective

The main objective of the research is to develop a methodology for a multidimensional analysis of energy poverty based on energy indicators and supplemented with non-energy data. The research problem focuses on creating an integrated set of indicators combining technical, economic, and social aspects, enabling a more precise diagnosis of the phenomenon than traditional single measures. The analysis covers the period 2018–2024.

Methods

The study uses data from the Household Budget Survey (HBS), conducted by Statistics Poland on a sample of approximately 32,000 households (2024 data). This source provides a representative database of information on expenditures, income, and living conditions, ensuring high reliability of the analysis. The study covers the entire surveyed population, which allows for drawing nationwide conclusions and identifying social and spatial disparities. Based on the source data, a set of energy poverty indicators was developed, reflecting both economic aspects and qualitative dimensions related to energy use and subjective assessment of housing conditions. The next stage involved an overlap analysis of indicators, enabling the identification of households affected by multidimensional energy poverty and the assessment of co-occurrence of different forms of energy deprivation. Subsequently, taxonomic methods were applied, including the Hellwig method, Weber method, and TOPSIS, as well as Principal Component Analysis (PCA). Non-energy data were incorporated into both approaches, enabling more precise classification of households, dimensionality reduction, and identification of the most relevant factors differentiating levels of energy poverty.

Results

The study showed that a change in the level of detail of questions in the HBS in 2020, affecting the LEAKS indicator, contributed to a noticeable decline in its national value. At the same time, observations from 2024 indicated an increase in selected non-financial indicators in regions affected by flooding, suggesting the sensitivity of these measures to sudden crisis events. The application of indicator overlap analysis enabled the visualization of the intensity of energy poverty. Taxonomic methods captured the direction and scale of changes in household conditions, while PCA enabled the identification of key factors influencing energy poverty levels. The results highlight the complexity of the phenomenon and importance of selecting appropriate analytical tools to assess its dynamics.

Conclusions

The developed synthetic index constitutes an innovative analytical tool enabling precise identification of areas requiring intervention in energy efficiency improvement. A key contribution to statistical methodology is the original combination of taxonomic methods with PCA, which enhanced the objectivity and stability of composite measures. The results indicate substantial regional disparities between voivodeships, which makes it necessary to analyse them individually rather than relying solely on national averages. This underscores the importance of a regional approach in diagnosing and monitoring energy poverty.

Keywords

energy poverty, taxonomy, principal component analysis (PCA), synthetic indicator

Objective

Within the economic literature, poverty is commonly conceptualized as a multidimensional phenomenon. One of these dimensions - in addition to the income dimension - is the housing situation of the population. The presentation aims to examine the relationship between income poverty and housing poverty across European countries, with particular emphasis on their co-occurrence and the extent of overlap between these dimensions in a comparative perspective.

Methods

The data used in the analysis come from the European Union – Statistics on Income and Living Conditions (EU-SILC) survey from 2020 for 30 European countries. EU-SILC provides harmonized and comparable microdata across European countries, making it a reliable source for analyzing social inequalities and living conditions. Its multidimensional approach, covering income, material deprivation, and housing conditions. The sizes of the datasets used is depend on the country and is e.g. 4192 for Cyprus and 19841 for Poland. The EU-SILC survey is essentially a representative survey, but the analyses used weights contained in the data set to correct deviations from the representativeness of the research sample.To identify housing poverty, the IFR (Integrated Fuzzy and Relative) method was used in a multidimensional non-monetary approach, taking into account technical, financial, and environmental dimensions. Monetarily poor households were identified using a poverty line defined as 60% of the median equivalized disposable income, calculated separately for each country. Next, the relationships between the risks of different types of poverty were examined using several measures of dependence, including Pearson`s linear correlation coefficient, Spearman`s rank correlation coefficient, distance correlation, mutual information, and normalized mutual information. Differences in housing poverty between monetary-poor and non-poor households were verified by a two-mean significance test.

Results

The results reveal substantial cross-country heterogeneity in both the level and structure of deprivation. The relationship between monetary and housing poverty is dimension-specific. A relatively strong and predominantly linear association is observed in the technical dimension, while the relationship in the financial dimension is weaker and less consistent. In contrast, the environmental dimension shows no clear relationship with monetary poverty. Correlation measures indicate near-zero dependence, and in some countries even suggest a reversed pattern, where higher-income households may experience greater environmental housing deprivation. An important result is the incomplete overlap between monetary and housing poverty – many households experience only one form of poverty.

Conclusions

The findings confirm that housing quality is only to a limited extent determined by household income and is instead strongly shaped by spatial, infrastructural, and policy-related factors. This points to the need for policy frameworks that move beyond income-based approaches and prioritize place-based interventions aimed at improving environmental quality and neighborhood conditions independently of redistribution mechanisms. Greater awareness of the risks associated with inadequate housing conditions can stimulate demand for high-quality, health-promoting housing and, consequently, strengthen public support for the development and implementation of policies that improve housing conditions.

Keywords

monetary poverty, housing poverty, multidimensional poverty measurement, fuzzy set approach (IFR), EU-SILC data

Objective

This study provides a comparative assessment of the fiscal costs and distributional effects of a stylised Job Guarantee (JG) across all 27 EU member states. The central research question is the extent to which the policy`s net cost and redistributive impact depend on the structure of the existing tax-benefit system and the welfare-state regime in place. The working hypothesis is that a structural tension exists: countries where a JG is fiscally cheapest are precisely those where its redistributive reach is weakest.

Methods

The analysis uses EUROMOD version J2.0+, the EU-wide tax-benefit microsimulation model maintained by the Joint Research Centre (JRC) of the European Commission. The micro-level data are EUROMOD`s harmonised input files derived from EU-SILC, conducted by Eurostat with national statistical offices on representative random samples of households. National sample sizes range from a few thousand households in smaller member states to roughly 25–30 thousand in the largest, totalling more than 250 thousand households and about 600 thousand individuals. Population estimates use harmonised grossing-up factors aligned with Eurostat conventions. The Job Guarantee is operationalised as a full-time public employment offer for the long-term unemployed, paid at 60 percent of the country-specific median wage. For each country a baseline and a reform scenario are simulated: the fiscal cost is decomposed into gross cost, benefit savings, additional tax and social-contribution revenue, and net cost. Redistributive effects are measured by the Gini coefficient, the at-risk-of-poverty rate at 60 percent of median equivalised disposable income, and decile income changes. Quality assessment relies on sensitivity analysis of the JG wage (80, 100, 120 percent of the benchmark) and on significance tests of cross-country correlations. Results are interpreted as conditional projections from a static microsimulation without behavioural responses.

Results

Preliminary results show pronounced cross-country heterogeneity in the effects of a Job Guarantee within the EU. Estimated net costs span a wide range - from a small fraction to several percent of GDP - with a median well below 1 percent of GDP. The self-financing ratio, defined as the share of the gross cost recouped through reduced benefit spending and additional tax and social-contribution revenue, also varies substantially across countries and is positively correlated with the generosity of the existing social protection system. Redistributive effects - poverty reduction, the change in the Gini coefficient, and income gains in the lowest deciles - are largest in countries with weaker benefit systems and smallest in countries with extensive welfare states.

Conclusions

The findings reveal a structural redistribution-cost tension: the same institutional features that make a Job Guarantee fiscally attractive also limit its redistributive reach. Economic assessment of the policy must therefore be conditional - universal claims about its `affordability` are misleading and should give way to analysis that takes the national tax-benefit context seriously. The study delivers the first coherent comparative assessment of a Job Guarantee in the EU together with a fiscal-cost decomposition along two absorption channels: benefit savings and additional tax and social-contribution revenue.

Keywords

Job Guarantee, microsimulation, EUROMOD, welfare state, redistribution

Objective

The aim of the study was to identify and compare the profiles of older adults across the Visegrad Group countries regarding health, material conditions, and economic security. The study sought to determine whether the situation of older adults is homogeneous or whether it exhibits diverse patterns of characteristics, and to what extent these profiles are similar across countries. Understanding the profiles of older adults enables us to assess whether similar indicator values across countries reflect genuine similarities in their situation.

Methods

The analysis utilised EU-SILC data (European Union Statistics on Income and Living Conditions) for 2024 on individuals aged 65 and older, comprising over 27,000 observations for Poland, the Czech Republic, Hungary, and Slovakia. The data come from a representative public statistics survey conducted on a random sample, and the use of cross-sectional weights allowed the results to be generalised to the elderly population in the analysed countries. In the first stage, a factor analysis based on a tetrachoric correlation matrix was conducted to identify the latent structure among variables describing health, material conditions, and economic security. Subsequently, using the Two-Step Cluster method, profiles of older adults’ functioning were identified. In the subsequent stage, a structural similarity analysis was applied, along with tests of the significance of structural differences and a procedure to identify distinctive differences between the obtained functioning profiles. The analysis was supplemented by an estimation of a multinomial logit model, allowing for the assessment of determinants of membership in specific profiles, taking into account demographic and spatial variables. All analyses were conducted using SPSS and Statistica.

Results

Four profiles of older adults’ functioning have been identified, the largest of which is intermediate, situated between the extreme configurations of deficits and resources. The results clearly indicate that health, material conditions, and economic security do not form a single coherent dimension, but rather a system of interdependent yet partially autonomous components. The structure of these profiles is generally consistent across countries, though the proportions and intensities of individual characteristics vary, with the greatest variation in health and functional limitations. At the same time, the observed profiles indicate the existence of mixed configurations, in which a favourable situation in one area coexists with deficits in other dimensions of functioning.

Conclusions

The results confirm that the situation of older adults is complex and heterogeneous, and that the observed functional profiles, while common across the V4 countries, take on different significance in different national contexts. This points to the limited usefulness of analyses based solely on aggregated indicators and underscores the importance of a structural approach. The results suggest the need to design social policy in a differentiated manner, taking into account both cross-national differences in the structure of the elderly population and the internal heterogeneity of the elderly population. Effective interventions require simultaneous consideration of multiple aspects of living.

Keywords

older adults, EU-SILC, economic security, profiling, identification of significant differences

Objective

The paper analyses the economic integration of Ukrainian immigrants in Poland, comparing pre-war labour migrants with refugees who arrived after the Russian invasion in 2022. It examines whether labour market outcomes improve with duration of stay and whether refugee integration follows different trajectories than earlier migration cohorts. The study also assesses selected dimensions of broader social integration, including language skills and housing independence

Methods

The study uses repeated cross-sectional data from five nationwide waves of surveys of Ukrainian immigrants conducted by the National Bank of Poland in 2019 and annually from 2022 to 2025. Each wave covered approximately 3,000–4,000 respondents and provides detailed information on labour market status, earnings, migration history, legal status, language proficiency and housing conditions. The analytical framework follows the international literature on immigrant labour market integration, focusing on convergence over time. Two complementary measures of migration duration are applied: years since the first non-touristic stay in Poland and years since the beginning of the current stay. These indicators allow identification of both accumulated migration experience and stability of residence. Key dependent variables include employment rates, average net wages (estimated from income categories), contract type, self-employment, self-assessed command of Polish, and housing independence. The study compares pre-2022 immigrants with post-2022 refugee cohorts, including persons under PESEL-UKR temporary protection. In addition to cross-sectional comparisons, synthetic cohort analysis is used by grouping respondents according to year of arrival and tracing cohort outcomes across subsequent survey waves.

Results

The results show clear differences between refugees and earlier immigrants. Pre-war migrants reached very high employment rates, often above those of the native population, while refugees started from much lower levels but integrated relatively quickly. Within three years, refugee employment exceeded 60%, faster than in many European countries. Wage gaps remained substantial: earlier migrants earned about PLN 1,000 more on average than post-2022 arrivals. Longer stay was also associated with a higher share of permanent contracts and some growth in self-employment. Language proficiency and independent housing improved strongly with time spent in Poland.

Conclusions

Poland’s temporary protection regulations implemented in Poland enabled unusually rapid labour market entry of Ukrainian refugees. However, convergence in wages and job quality is slower than convergence in employment. Time spent in Poland remains a key driver of integration, not only economically but also socially through language acquisition and housing independence. The study contributes to migration research by combining cohort and cross-sectional approaches and offers evidence relevant for labour market, housing and integration policy design.

Keywords

Ukrainian migrants, refugees, labour market integration, Poland, migration cohorts

Objective

The main research question addressed in this article concerns whether and how the returns to education and work experience in Poland have changed over the last twenty years, and whether these changes vary according to gender, educational attainment and work seniority. The authors aim to determine whether traditional human capital factors continue to play a key role in shaping wages, or whether their significance is diminishing in the context of structural and technological changes in the economy.

Methods

This study applies a quantitative, longitudinal research design to estimate returns to education and work experience in Poland over 2004–2024. The empirical framework is based on an extended Mincer earnings function, where the logarithm of wages is explained by education level, work experience, its quadratic term, and time effects. The model is estimated on pooled cross-sectional data with time dummies, enabling the analysis of both average returns and their changes over time. Econometric regression analysis constitutes the main analytical tool, allowing estimation of marginal effects of education and experience. Additional specifications include gender-specific models and variants capturing heterogeneity across education and experience levels. Robustness checks involve testing model stability and including macroeconomic controls such as the unemployment rate. The analysis uses secondary data from the Structure of Earnings Survey conducted by Statistics Poland (GUS). The dataset is aggregated at the three-digit occupational level and covers firms employing at least ten workers, ensuring high reliability of wage data. The sample includes 13,914 observations across 11 waves. Education is assigned to occupations based on ISCO classification, adjusted to the Polish context, while experience is measured in intervals.

Results

The results indicate a decline in the rates of return on both education and work experience. The average rate of return on education fell from 17.2% in 2004 to 15.5% in 2024, with this decline being particularly pronounced among women. Significantly, a reversal in gender relations occurred in 2018. From 2018 onwards, women began to achieve lower returns on education than men. A similar downward trend applies to work experience, where the premium fell from 16.1% to 11.2%, and the decline was even more pronounced for women. An analysis by level of education shows that although the highest education premiums relate to higher education, a decline is being observed. At the same time, returns on vocational education remain relatively higher compared to general education at secondary level.

Conclusions

The findings of the study point to a significant transformation in the mechanisms shaping wages. The declining importance of education and experience suggests that traditional measures of human capital are losing their significance. The results support the hypothesis that the expansion of higher education leads to a compression of the wage distribution. From an economic policy perspective, this implies a need for greater balance in the education system, in particular the development of lifelong learning and the promotion of technical education. At the same time, the observed changes in gender differences indicate that equal access to education does not translate into equal outcomes in the la

Keywords

Mincer equation, returns to education, work experience, gender differences