Day 2 (Room C, July 2) - The 5th Congress of Polish Statistics

Objective

The aim of the study is to examine the application of non-random data sources in business statistics.

Methods

The study uses the simulation method on the basis of a superpopulation model built from the frame and dataset of a real business survey.

Results

The study shows a comparison of the accuracy of the used estimators with the unbiased Horvitz-Thompson estimator.

Conclusions

Estimators using mass estimation are feasible in terms of the acceptable levels of accuracy compared to the representative method.

Keywords

mass imputation

Objective

The aim of the study is to present an analysis of the activities carried out to develop PKD 2025 and the promotion of the classification in order to encourage domestic users to independently re-register their PKD activity codes in official registers, including REGON, KRS and CEiDG.

Methods

The study shows the statistics and analysis of the impact of the wide-ranging national consultations on the PCAs in the new classification. The changes introduced, in particular regarding the divisions of the current groupings, should be duly justified (for example, turnover or a significant estimated market share).

Results

As a result of meetings organized by Statistics Poland and promotion of changes to the PKD 2007 in the years 2022-2024, 71 institutions, chambers, business associations, government administration units and the UAS applied. The largest group of entities that submitted proposals for changes to the classification were institutes, chambers and associations of entrepreneurs. As part of a broad consultation on the amendment to the PKD 2007, 173 proposals for changes to the classification were collected. In addition, as part of the legislative process, further changes and new c

Conclusions

Classification of PKD activities is a classification with a wide application, used, among others, for reporting, registration and economic analysis purposes. For these reasons, it is important that it reflects the changing reality, i.e. it takes into account new technologies, new activities emerging on the market, new categories of service activities.

Keywords

PKD 2025, PKD amendment, new PKD, classifications

Objective

• Practical use of the τ-Argus program on aggregated data.
• The rules of Statistical Disclosure Control in DG-1 Survey.
• The importance of statistical suppression in data protection.

Methods

Statistical disclosure control techniques can be defined as set of methods to reduce the risk of disclosing information on individuals, businesses or other organisations. SDC methods minimise the risk of disclosure to an acceptable level while releasing as much information as possible.
Statistical disclosure control primary and secondary suppression with the program τ-Argus.
We use these methods in the DG-1 survey: 'Assessment of the current economic activity of enterprises' (completed on the DG-1 form, 'Report on economic activity').
This is one of the main sources of information on the current economic situation of entities in Poland. The data obtained reflect the economic and social situation in the country and each of its provinces.
Form DG-1 is a monthly survey which includes legal entities, organisational units without legal personality, and natural persons with 10 or more employees whose activities have been classified according to the 2007 Polish Classification of Economic Activities (PKD) as falling within the scope of selected sections.

Results

Preparation of files needed for statistical disclosure control with the τ-Argus program.
Practical presentation of the steps statistical disclosure control primary and secondary suppression in τ-Argus program for single data files and related tables.

Conclusions

Since January 2016, statistical disclosure control in the DG-1 survey has been determined based on the value of the cumulative turnover variable (On). This means that the occurrence of dominance in the aggregate is determined by the variable On. The τ-Argus programme is used to determine which aggregates are covered by primary and secondary suppression.
We also use the τ-Argus program to analyze statistical suppression in other tables and surveys.
The τ-Argus program simplifies statistical disclosure control, especially secondary suppression.

Keywords

short-term statistics, DG-1 survey, statistical disclosure control, τ-Argus

Objective

This study aims to assess the financial effects of internationalization in the food industry by comparing exporters and non-exporters (using ROA, ROE and ROS), and by analyzing the impact of internationalization intensity (foreign revenue share) on these metrics. Hypotheses are divided into two sets: basic export status comparison and continuous level of internationalization.

Methods

The study applies fixed-effects panel regression, complemented by Hausman, Breusch-Pagan and robust standard error corrections (Huber-White). The dataset includes over 9,800 firm-year observations from 1998 to 2023. Dependent variables are log-transformed ROA, ROE and ROS, and independent variables include internationalization level, revenue, total assets and market capitalization.

Results

No significant differences in ROA, ROE or ROS were found between exporters and non-exporters. However, a higher level of internationalization significantly reduced ROA and ROS, with ROE remaining statistically unaffected. For ROS, a nonlinear, escalating negative effect was detected.

Conclusions

Internationalization does not automatically improve profitability in the food industry. Negative effects on ROA and ROS may reflect expansion costs, operational complexity and delayed returns. This study challenges common assumptions and contributes robust empirical evidence to the literature on firm internationalization.

Keywords

internationalization, financial performance, food industry, panel data, ROA, ROE, ROS

Objective

In the recent years, there has been a dynamic, steadily deepening decline in the population of Poland. This process is shaped primarily by natural increase, which has remained negative since 2013. A record high natural decline was observed in the past 2024, when 156.7 thousand more people died than were born (a decline of 41 people per 10 thousand population). The aim of this paper is to present the most important demographic factors influencing unfavorable trends in mortality and birth rates.

Methods

The source of the data are the results of demographic surveys conducted by the Statistics Poland – collected by the Statistical Office in Olsztyn. The characteristics of population reproduction processes were characterized by the following coefficients: demographic dynamics, natural increase and fertility. The observed trends in the intensity of births and deaths were referred primarily to the transformations in the population structure by age. Various forms of graphic presentation were used to visualize the analysis results, such as: cartograms, age pyramids and charts.

Results

The dynamic decline in the number of births (apart from the procreative behaviors of the population) is influenced by the decreasing number of women of reproductive age, especially those aged 20-35. In the case of deaths, the greatest impact on their intensity is the increasing percentage of people aged 65 and over in the total population (the ongoing process of ageing of population). Additionally, in 2020 and 2021, the covid-19 pandemic had a particular impact on the rapid increase in the number of deaths. Analyses also showed clear spatial differentiation in terms of vital statistics.

Conclusions

The observed demographic processes may significantly affect all areas of socio-economic life in Poland. The demographic forecast indicates that the current trends regarding natural decrease and ageing of population will deepen.

Keywords

natural increase, depopulation, vital statistics, ageing of population

Objective

The aim of this presentation is to introduce capture-recapture methods for population size estimation (dual-system estimation, multiple-system estimation) that account for record linkage errors and their implementation in the {uncounted} package in R. The package is currently being developed at the Population Research Methodology Center in the Statistical Office in Poznań.

Methods

The presented tool includes classical estimators as well as estimators that incorporate record linkage error correction (one-way, symmetric, and asymmetric (Di Consiglio * Tuoto, 2018: Zult et al., 2025)).

Results

The package offers the capability for variance estimation and construction of confidence intervals to assess the quality of the estimates.

Conclusions

As part of the presentation, we will showcase the package in the context of administrative data, with special emphasis on modeling record linkage errors. Examples with simulated data will illustrate the package`s capabilities in various scenarios.

Keywords

capture-recapture, population size estimation, administrative registers, record linkage errors

Objective

Development and release of a universal, error-tolerant tool (package) for record blocking that minimizes the number of comparisons in entity resolution procedures without sacrificing matching accuracy, especially under conditions of missing identifiers and low data quality.

Methods

Approximate Nearest Neighbours, Graph indexing, record blocking

Results

Development of a package that provides high recall while achieving a high reduction of pairs required for comparison.

Conclusions

The package is a tool suitable for use in integrating data from administrative registers for the purpose of population size estimation. A scientific article describing the proposed method is currently under review in the journal SoftwareX.

Keywords

record linkage, deduplication, approximate nearest neighbours, official statistics

Objective

Identification of further residence plans of Ukrainian citizens residing in the Łódź and Silesian voivodeships, i.e. in voivodeships with a diversified economic context.

Methods

A logistic regression model was used to analyze the further residence plans of Ukrainians in the Łódź and Silesian regions. The explanatory variables included, among others, those illustrating the situation of immigrants on the labor market, knowledge of the Polish language, their satisfaction with selected areas of life (financial, housing situation) and the type of support received. The main source of data is a survey conducted among Ukrainians living in the Łódź and Silesian voivodeships in the fall of 2024. The tool used was an online survey - in total, the survey covered 680 Ukrainians.

Results

The results obtained indicate significant differentiation in terms of further residence plans, both among migrants who arrived before and after the war, as well as between the analysed voivodeships. The results of the analysis indicate, among other things, that knowledge of Polish increases the probability of declaring a desire to stay in Poland, especially in the case of war refugees. People who declared satisfaction with their financial situation expressed over 1.6 times greater willingness to stay in the Łódź region, and 1.4 times greater in the Silesia region.

Conclusions

The conducted study makes it possible to estimate the number of Ukrainians who are willing to stay permanently in selected regions and the factors that increase the chance of staying in the region they currently live in.

Keywords

Ukrainians, labor market, Lodz Voivodeship, Silesian Voivodeship

Objective

The study presents the demographic processes of the Łódź and Świętokrzyskie voivodeships in 2002-2022. The following research goals were set: analysis of changes in the population of counties and municipalities in 2002-2022: identification of areas with annual population loss in the period under study: identification of factors of population change in these areas (typology): analysis of the population ageing (including classification of municipalities in terms of similarity in changes in the age structure of the population).

Methods

In the study selected descriptive statistics tools were used, as well as selected cluster analysis methods. The classification of counties according to population change determinants was also carried out. The source of statistical data were the publications of the Central Statistical Office for counties and municipalities of the Łódź and Świętokrzyskie voivodeships in 2002–2022.

Results

The obtained results illustrate the demographic problems of the Łódź and Świętokrzyskie voivodeships, characterized by a relatively high intensity of the depopulation process (in many municipalities as a result of both natural and migration decrease) and a high level of demographic aging. Particular attention was paid to counties and municipalities characterized by annual population loss in the period 2002-2022.

Conclusions

The study provides a synthetic picture of the demographic situation in the central macroregion over two decades. The obtained results identify areas of population growth and areas of the greatest, often long-term demographic decline. The main factor of population changes in the counties and municipalities of the central macroregion is natural decrease. Migration-related population loss is of particular importance in the depopulation process in the case of unattractive areas, especially those located on the outskirts of voivodeships.

Keywords

depopulation, ageing process, determinants of population changes, central macroregion,