Day 1 (Room D, July 1) - Congress Replay - The 6th Congress of Polish Statistics

Objective

The aim of the study is to explain the sharp fertility decline in Poland, Lithuania, Latvia, and Estonia between 2017 and 2024 by distinguishing between structural and behavioral determinants. The analysis seeks to assess to what extent the observed changes result from demographic shifts versus behavioral responses to overlapping crises—namely the COVID-19 pandemic and the war in Ukraine—and the associated economic and social uncertainty.

Methods

The study is based on demographic data for the period 2017–2024, including vital statistics, Eurostat data, and national statistical office records. The core methodological approach is the decomposition of changes in the total fertility rate (TFR) using the frameworks developed by Kitagawa (1955) and Das Gupta (1993), which allow the separation of structural components (changes in the age and parity composition of women) from behavioral components (changes in fertility rates within these groups). Additionally, tempo-adjusted fertility measures following Bongaarts and Feeney (1998) are applied to correct for distortions caused by the postponement of births. The analysis includes age-specific and parity-specific fertility rates, enabling the identification of shifts in first, second, and higher-order births. To evaluate the impact of crises, a counterfactual event-study design is employed, comparing post-2022 fertility trajectories with pre-pandemic trends and with selected Western and Central European countries that differ in institutional stability and proximity to the conflict. This approach allows capturing both the direct effects of shocks and their interaction with longer-term demographic processes.

Results

In Poland, the total fertility rate declined from approximately 1.45 in 2017 to around 1.10 in 2024, representing a drop of nearly 24%. Decomposition results indicate that about 80% of this decline was driven by behavioral factors, while only 20% can be attributed to structural changes. The most pronounced decreases were observed among women aged 25–34, particularly in first births, which account for the majority of the post-2021 decline. At the same time, the mean age at childbirth increased, indicating postponement of reproductive decisions. The combined impact of the pandemic and the war produced a “double shock” that reinforced lasting changes in fertility behavior.

Conclusions

The findings suggest that the fertility decline in Poland is primarily behavioral and reflects responses to rising uncertainty caused by the pandemic and the war in Ukraine, rather than long-term structural demographic changes. The postponement of first births plays a central role and may have lasting implications for future fertility levels. The study highlights the importance of economic and geopolitical factors in shaping reproductive behavior and points to limited demographic resilience in the face of compounded crises.

Keywords

fertility, uncertainty, COVID-19 pandemic, war in Ukraine, demographic decomposition

Objective

The aim of this paper is to present the findings of the European survey on gender-based violence (EU-GBV ) in Poland, as well as to address the scale, structure as well as variations of this phenomenon depending on gender and social context. The analysis focuses on identifying forms of violence and differences in how they are perceived and reported by respondents.The use of a common methodology across European Union countries makes it possible to obtain comparable results and assess the phenomenon of violence.

Methods

The EU-GBV survey was carried out as part of a project coordinated at European Union level and funded by Eurostat. In Poland, the survey was conducted by the Statistics Poland. As part of the project, a survey on gender-based violence (EU-GBV) was conducted at national level on a randomly selected sample of respondents (women only), using a questionnaire and methodology developed at European level. The data come from a representative survey conducted on a sample of adults (aged 18 and over), selected at random, allowing the results to be generalised to the country’s population. A standardised questionnaire covering various forms of violence (physical, psychological, sexual, economic and digital violence) was used as the basis for developing the Polish version of the questionnaire. The analysis utilised statistical weights to adjust the sample structure relative to the population. The quality of the results was assessed through an analysis of estimation errors and the application of basic statistical inference methods, including confidence intervals. Limitations arising from the underreporting of violence and differences in respondents’ willingness to disclose their experiences were also taken into account.

Results

The results indicate a significant scale of gender-based violence (EU-GBV) in Poland and its clear gender, age and context differences. The survey was conducted among a group of randomly selected female respondents across sixteen provinces in Poland. Most often, forms of psychological and economic violence are reported, with women more often declaring experience of violence in private relationships, and men less often revealing such events. There has also been a significant contribution of violence in the digital space and in the workplace. Interpretation of the results indicates the impact of social and cultural norms on the level of disclosure of violence.

Conclusions

The obtained results underline the need to further develop research on gender-based violence using methods of official statistics and to improve measurement tools. It is important to take into account differences in reporting violence and strengthen procedures to ensure the reliability of data. The results are important for shaping public policies, including preventive and intervention measures, and contribute to the development of methods for the analysis of social phenomena of a sensitive nature.

Keywords

gender-based violence, violence against women, mobbing, stalking, childhood violence

Objective

The study aims to examine how differences in the institutional design of long-term care (LTC) in Central and Eastern Europe, particularly in Poland and Czechia, are associated with older adults’ quality of life. Special attention is given to two regional LTC trajectories: the more family-based and subsidiarity-oriented Polish model and the Czech model of limited universalism. These are further contrasted with LTC regimes in the Netherlands, Germany, and Italy.

Methods

The analysis uses representative panel data from the SHARE study (Survey of Health, Ageing and Retirement in Europe), covering 28,025 older adults across five waves in Poland, Czechia, Germany, the Netherlands, and Italy. Country selection was purposeful and comparative. Poland and the Czech Republic are treated as two Central and Eastern European cases that, despite sharing a common regional classification, represent distinct institutional arrangements for long-term care. The Netherlands, Germany, and Italy, representing distinct LTC regimes: Nordic, mixed, and family, provide a comparative context. Quality of life is measured using the CASP index and its domains (control, autonomy, self-realization, pleasure), along with overall life satisfaction. Key explanatory variables include four LTC dimensions: formal institutional care, formal home care, receiving help, and providing help. Models control for age, gender, physical and mental health, financial situation, education, marital status, GDP per capita, income inequality, and the COVID-19 period, treated as a natural experiment (“stress test”) for LTC systems. Regression mixed models were applied to analyze the overall relationships between LTC forms and quality of life and their variation across countries.

Results

Receiving help is associated with lower quality of life, particularly in control and self-realization, with stronger effects in Poland and Czechia than in Germany and the Netherlands, suggesting a higher burden of dependency where LTC relies more on family or has weaker formal support. Formal home care is also linked to lower CASP, especially in control. Institutional care shows no clear negative association after controls. A notable contrast emerges: providing help is positively associated with quality of life in Poland but negatively in Czechia. The strongest predictors remain physical and mental health and perceived financial situation.

Conclusions

Central and Eastern Europe is not homogeneous in how LTC relates to older adults’ quality of life. Poland and Czechia, despite similar regional classifications, exhibit distinct patterns of relationships between receiving and providing care and the well-being of older adults, indicating that LTC outcomes depend not only on formal care types but also on institutional, familial, and cultural contexts, and on whether help is experienced as support or dependence. Similar LTC forms may thus have different consequences across care regimes.

Keywords

long-term care: older adults’ quality of life: Poland: Czechia: SHARE

Objective

The aim of this study is to examine the level and heterogeneity of job satisfaction among workers aged 50 and above in Europe in the post-COVID-19 period. The paper investigates whether disruptions to work during the pandemic, such as remote working and changes in working time, are associated with job satisfaction and how patterns of satisfaction differ across countries and socio-economic characteristics of older workers.

Methods

The analysis is based on data from Wave 9 (2021–2022) of the Survey of Health, Ageing and Retirement in Europe (SHARE), a cross-national, representative survey covering more than 20 European countries and Israel. The analytical sample includes 4,858 respondents aged 50 and above who were economically active at the time of the interview. Job satisfaction is assessed using a multidimensional set of indicators describing working conditions, including physical workload, time pressure, autonomy at work, job security, opportunities for skills development, recognition and support, as well as the perceived adequacy of earnings. All variables are harmonised across countries following SHARE methodological guidelines. In the first step, exploratory factor analysis is applied to identify latent dimensions underlying job satisfaction and to reduce the dimensionality of the observed indicators. Factor scores are subsequently used as inputs in a cluster analysis, which allows the identification of distinct and internally homogeneous job satisfaction profiles among older workers. In the final step, multinomial logistic regression models are estimated to examine the associations between cluster membership and individual characteristics such as age, gender, and educational attainment, as well as pandemic-related work disruptions, including the experience of remote working and changes in working hours during the COVID-19 period.

Results

Four job satisfaction profiles are identified: physically strained workers (26.8%), dissatisfied workers (27.0%), discouraged workers (25.2%), and satisfied workers (20.9%). Thus, around 80% of workers aged 50 and over experience limited job satisfaction, although for different reasons. The distribution of these profiles varies markedly across countries. Regression results indicate that pandemic-related work disruptions had a limited overall impact on job satisfaction, while remote working reduced the likelihood of belonging to the physically strained group. Higher educational attainment and greater perceived job security significantly increase the probability of being satisfied with one’s job.

Conclusions

The findings suggest that in the post-pandemic period the key challenge for extending working lives lies in job quality rather than in temporary pandemic-related disruptions. Low job satisfaction among older workers is widespread and strongly differentiated across countries. Policies aimed at improving working conditions, especially by reducing physical strain and strengthening job security and recognition, may play an important role in promoting longer and more sustainable employment among people aged 50 and above.

Keywords

job satisfaction: older workers: COVID‑19: SHARE: labour market

Objective

The study examines economic resilience from a micro-spatial perspective, focusing on how local industrial structures and firm dynamics shape variation within regions. Resilience is defined through four components: resistance, recovery, renewal, and reorientation. The paper introduces a grid-level measure - the Local Industrial Resilience Index (LIRI), to identify neighbourhood-level patterns and assess how diversity, specialisation and related variety are associated with local resilience.

Methods

The analysis is based on geo-referenced firm-level data from the REGON register for the Mazowieckie region (Poland), covering two cross-sections: 2012 and 2021. The study uses the full population of firms (approximately 989,000 in 2012 and 1.11 million in 2021), rather than a sample, which eliminates sampling-related uncertainty and the need for statistical inference based on sampling error. Firms are assigned to a regular 2×2 km grid (approximately 9,320 cells), enabling a fine-grained spatial analysis. The study introduces LIRI, a composite measure integrating four dimensions of resilience: resistance, recovery, renewal, and reorientation. The index combines information on firm survival and entry with indicators of local industrial structure, including diversity, specialisation, concentration, and related variety. Resistance and recovery are estimated as standardised residuals from econometric models (binomial model for firm survival and negative binomial model for firm births), comparing observed and expected values. Renewal and reorientation capture entries into related and unrelated sectors, based on measures of sectoral relatedness derived from co-occurrence patterns. This approach allows controlling for structural characteristics and identifying deviations from expected dynamics at the local level.

Results

The results show that economic resilience varies significantly across space and does not follow a simple core-periphery pattern. While the Warsaw metropolitan core exhibits high resilience, several medium-sized towns also demonstrate strong adaptive capacity. In contrast, many peripheral and mono-industrial areas remain structurally vulnerable. Related variety emerges as a key factor supporting both firm survival and renewal, whereas high sectoral concentration is associated with weaker resilience across all dimensions. These findings highlight the importance of local industrial structure in shaping resilience outcomes.

Conclusions

The results indicate that regional resilience is shaped by local industrial structures rather than scale alone. Areas characterised by greater diversity and related variety show higher adaptive capacity, while sectoral concentration increases vulnerability to shocks. The proposed LIRI measure provides a consistent framework for analysing resilience at a fine spatial scale. The findings contribute to empirical research on resilience and offer useful insights for place-based regional policy, particularly in supporting diversification and strengthening local economic linkages.

Keywords

composite measure: local resilience: firm resistance: firm renewal: spatial analysis

Objective

Categorical variables such as occupation (ISCO) or industry (NACE) classifications frequently change their encoding between survey waves, which precludes longitudinal comparisons. This study presents the cat2cat method for harmonizing inconsistently coded categorical variables and demonstrates its empirical application to the analysis of the relationship between age structure and wages across occupational groups using Polish Structure of Wages and Salaries Survey data (2006–2014).

Methods

The cat2cat method was originally proposed by Nasiński, Majchrowska, and Broniatowska (2020), subsequently formalized and extended by Nasiński and Gajowniczek (2023). The method has been further refined following expert review, including improvements in error handling and diagnostic transparency. The cat2cat algorithm uses a transition table for harmonization. The mapping is bidirectional, enabling harmonization from old encoding to new and vice versa. When encoding results from classification evolution, mapping of the new (more detailed) encoding to the old (less detailed) one is often ambiguous, whereas the reverse mapping may be more straightforward. The 2010 revision of the Polish Classification of Occupations (KZIS) affected approximately 85% of 4-digit codes, previously precluding the analysis of wages across specific occupational groups over time. Using the cat2cat method, comparability of SWZ data for 2006–2014 was restored and an augmented Mincer wage equation was estimated at the 3-digit occupational group level. The variable of interest is the share of workers aged 55–65 in a given occupational group. The empirical application uses 5 waves of the Structure of Wages and Salaries Survey (SWZ, 2006–2014, approx. 700,000 observations per wave). The design weight variable from SWZ was used to generalize results to the population.

Results

In the full-sample regression, the age structure coefficient changed significantly over time: before 2010 it was negative or near zero, whereas after 2010 it became positive, potentially indicating the decline of a *seniority premium* characteristic of the earlier period. At the aggregate level for Professionals, an unexpectedly strong negative correlation between wages and the share of older workers was observed. Disaggregation to the 3-digit level, however, revealed considerable internal variation — for example, wages of older physicians increase with experience, whereas the knowledge of older engineers and IT specialists may become obsolete.

Conclusions

The cat2cat method, progressively developed from an initial concept (2020) through formalization and open-source implementation (2023, R / Python) to further refinements informed by expert review, provides a complete tool for harmonizing evolving categorical classifications in longitudinal data. Its application to Polish wage data demonstrates both the practical utility of the method and novel findings on the heterogeneity of the relationship between age structure and wages across occupational groups. The methodology is applicable wherever categorical encodings change over time, including epidemiology (ICD) and economic activity statistics (NACE).

Keywords

data harmonization, categorical variables, longitudinal studies, machine learning, statistical software

Objective

The aim of the presentation is to present a methodological approach to using data from online job advertisements as a non-statistical source of labor market information in the study of skills requirement. In particular, we will discuss how the data were processed and analyzed. Additionally, the objective is to demonstrate how a properly designed methodological approach can enhance the usefulness of job advertisement data in labor market research, while taking into account limitations related to data quality, representativeness, and the specific nature of this type of source.

Methods

In the described model, a research approach combining qualitative and quantitative analysis was applied in order to comprehensively capture the structure and dynamics of skills requirement in selected occupations based on online job advertisements. The adopted perspective assumes the use of complementary analytical techniques, which increases the validity of inference and reduces the limitations arising from the specific nature of the data source. In the presented research approach, exploratory methods were used in the qualitative analysis, including content analysis and approaches specific to natural language processing (NLP), enabling, among others, the identification and standardization of skill category names. Categorization procedures were applied to develop a skills grouping framework, leading to the creation of a structured, original classification of skills used in further analysis. In the quantitative analysis, statistical description and elements of machine learning were employed, including clustering techniques and co-occurrence analysis, which made it possible to test an alternative method for identifying key skills in specific occupations. The data sources included information from online portals containing job advertisements. The analysis was conducted on a purposive sample of online job advertisements for selected occupations, within which more than 26 million instances of required skills were analyzed over the period from 2018 to 2025.

Results

The skills analytical model developed as part of this work is an example of how a properly designed methodological approach can increase the usefulness of job advertisement data in labour market research, while taking into account limitations related to data quality, representativeness, and the specific nature of this type of source. The OJA database as a data source can supplement analyses of labour demand and labour market changes, contributing to a better alignment between education and the labour market. It can also support the identification of skills gaps and supplement information derived from statistical data.

Conclusions

The skills analytical model developed as part of this work is an example of how a properly designed methodological approach can enhance the usefulness of job advertisement data in labor market research, while taking into account limitations related to data quality, representativeness, and the specific nature of such sources. The OJA database as a data source can support analyses of labour demand and labour market changes, contributing to a better alignment between education and labor market needs. It can also support the identification of skills gaps and supplement information derived from statistical data.

Keywords

online job advertisements, skills demand, new data sources, web scraping, ISCO

Objective

This study aims to provide a comprehensive global validation and extension of the entropy-marginal product model introduced by Bwanakare, Cierpiał-Wolan, and Rzeczkowski (2025), confirming their groundbreaking theoretical hypothesis that CO2 emissions are fundamentally driven by the interaction between thermodynamic entropy efficiency and energy marginal productivity. The research seeks to establish universal applicability of this framework across diverse economic systems and development stages, while demonstrating its superior predictive power compared to conventional energy intensity metrics

Methods

The study analyzes 127 countries over 23 years (2000-2022), representing 98.4% of global CO2 emissions. The theoretical framework utilizes nonextensive entropy based on Tsallis statistics with parameter (q), opposing conventional Boltzmann-Gibbs statistics. This aligns with Bwanakare`s theoretical predictions that economic-energy systems exhibit long-range dependencies inadequately captured by extensive thermodynamic frameworks. The entropy-to-marginal product ratio (?) serves as the key analytical metric. Analytical methods include spatial econometrics identifying cross-country spillover effects, panel regression with fixed and random effects controlling for heterogeneity, and dynamic modeling capturing temporal ? evolution. The study analyzes ? heterogeneity across developed economies, emerging markets, and developing nations. Comparative analysis evaluates explanatory power between the entropy-marginal product framework and conventional indicators including energy-to-GDP ratios and carbon intensity metrics. The research incorporates out-of-sample forecasting, temporal trend analysis, and optimization modeling for climate finance allocation. Robustness checks include sensitivity analysis for parameter q and cross-validation procedures ensuring reliability across contexts.

Results

The results confirm the main hypothesis, demonstrating that the ? ratio achieves explanatory power of R2 ranging from 0.847 to 0.923, surpassing conventional indicators (R2 = 0.521-0.687) with predictive improvement of 35-62%. Systematic heterogeneity of ? was documented: developed economies 0.847, emerging markets 0.789, developing countries 0.712. Temporal reinforcement amounts to +36% since 2000, with acceleration of +16.6% after the Paris Agreement. Spatial analysis reveals significant spillover effects (? = 0.156, p * 0.001). Entropy-optimized finance allocation achieves 26.3% higher cost-efficiency than GDP-based approaches. The parameter q exceeds unity (mean: 1.21, range: 1.12-1.31), confirming the necessity of Tsallis statistics.

Conclusions

The study represents a breakthrough in understanding CO2 emission mechanisms, empirically demonstrating the necessity of describing economic-energy systems within the framework of nonextensive thermodynamics. The entropy-marginal product metric enables identification of high-impact intervention points and optimization of resource allocation. The demonstrated 26.3% higher cost-efficiency suggests that policymakers should adopt the nonextensive entropy framework as the foundation for climate policy. Identified cross-country spillover effects underscore the importance of international coordination, while systematic ? heterogeneity indicates the need to differentiate policies according to countr

Keywords

CO₂ emissions forecasting: Tsallis entropy: marginal product of energy: thermodynamic efficiency: panel econometrics