Journal of Social and Economic Statistics Feed

Quantifying the Value of Fine-Tuning: A Reproducible Framework for Financial Sentiment Analysis in the Technology Sector

Mon, 29 Jun 2026 00:00:00 GMT

Financial sentiment analysis in the technology sector remains difficult because market-moving news combines firm-specific events, macro signals, and highly technical vocabulary. This study evaluates whether task-specific fine-tuning improves on zero-shot use of FinBERT when the corpus is restricted to English-language Technology news collected from MarketAux between 1 January and 30 June 2025. The final gold set contains 894 manually reviewed headline-description pairs labeled as negative, neutral, or positive and split into 625 training observations, 134 validation observations, and 135 test observations. On the held-out test set, zero-shot FinBERT reaches an accuracy of 0.637 and a macro F1 score of 0.626, whereas the fine-tuned version reaches 0.585 and 0.532. An exact McNemar test does not indicate a statistically significant difference between the two classifiers (p = 0.450). The findings suggest that, under a small sector-focused design and consumer-grade computational constraints, fine-tuning is not automatically superior to a strong domain-pretrained baseline.

AI Techniques for Survey Data Quality: Transformers and GANs

Mon, 29 Jun 2026 00:00:00 GMT

This study develops a comparative framework for missing data imputation in social surveys, integrating classical statistical methods, machine learning algorithms, and advanced deep learning methods, including Transformers and Generative Adversarial Networks (GANs). The analysis specifically addresses the imputation of categorical and ordinal variables, ensuring full alignment between imputation strategies and classification-based evaluation metrics, thereby overcoming a key limitation of existing approaches. An empirical evaluation on real data from the Italian “Aspects of Daily Life” (AVQ) survey—covering approximately 50,000 individuals and 735 variables—demonstrates that deep learning models, particularly Transformer-based architectures, consistently outperform traditional and machine learning methods in terms of predictive accuracy, generalization, and preservation of distributional properties. The findings show that attention-based models effectively capture complex dependencies in high-dimensional survey data and can improve the quality and reliability of official statistics.

Fertility Patterns and Regional Disparities in Romania: Evidence from Spatial Panel Models

Mon, 29 Jun 2026 00:00:00 GMT

Romania, like many other European Union countries, has been facing a persistent decline in fertility, raising significant demographic and socio-economic concerns. Understanding regional disparities is essential for developing effective demographic policies. This paper examines fertility differences across Romanian counties (NUTS 3) between 2013 and 2023, using spatial econometric and panel data approaches. Exploratory Spatial Data Analysis (ESDA) revealed clear regional patterns: the North-East and Central regions (especially Suceava, Botoșani, Bistrița-Năsăud, Harghita, and Mureș) consistently registered high fertility rates, forming high–high clusters, while the South-West and West regions (notably Gorj, Vâlcea, and Hunedoara) exhibited persistently low fertility, forming low–low clusters. The Moran’s I index, calculated for each year, confirmed a significant positive spatial autocorrelation (p < 0.01) throughout the entire period, with a slight decline during 2020-2021, potentially reflecting the impact of the COVID-19 pandemic. The estimated SARAR model with fixed effects confirmed the spatial dependence of fertility across counties. Results indicate that regional macroeconomic wealth (GDP per capita) is not statistically significant, whereas higher female labour force participation positively impacts fertility. Conversely, a high reliance on social assistance, high crude emigration rates, and significant educational burdens depress birth rates, revealing complex socio-economic dynamics. These results indicate that fertility patterns are shaped not only by structural and cultural factors but also by spatial spillovers. The findings highlight the importance of incorporating spatial dependence into demographic research and support the design of targeted, region-specific policies to reduce fertility disparities in Romania.

The Economic and Social Determinants of NEET Rates in European Union Countries

Mon, 29 Jun 2026 00:00:00 GMT

The transformations of the European labour market have increased the relevance of studies addressing occupational vulnerability and the challenges associated with the professional integration of young people. This paper aims to examine the economic and social determinants influencing the NEET (Not in Education, Employment, or Training) rate across the European Union, while also identifying the structural differences among European economies according to labour market characteristics. In order to achieve this objective, we use two panel data regression models in order to analyze the effects of economic, occupational, and social factors on the NEET rate, and to highlight the structural differences identified between genders. The results of the analysis show that the female NEET rate does not appear to be driven primarily by labour market fluctuations or by conventional macroeconomic factors, but rather by the continuity and stability of women’s participation in economic activity. For the young men, our model identifies a broader set of factors associated with occupational vulnerability, dominated by long-term unemployment, poverty risk, and the persistence of unemployment spells as the main exclusionary mechanisms.

Factors Influencing the Decision to Join a Pension System. Examination of Youth Perspectives in Colombia

Mon, 29 Jun 2026 00:00:00 GMT

Over the past 30 years, Colombia’s pension systems have undergone substantial changes, leading to the establishment of two primary systems: the Solidarity Average Premium Scheme and the Individual Savings Scheme with Solidarity. This study aimed to discover and assess the factors influencing Colombian youth’s decision to join a pension fund, a subject that has been scarcely examined in the Colombian context. A digital questionnaire was developed utilising the Theory of Planned Behaviour and completed by 482 youths, whose data were evaluated using a Structural Equations Model. The findings indicate that attitudes towards pension funds, beliefs, social influences, and perceived behavioural control favourably impact juvenile decision-making. Furthermore, we discovered that political uncertainty, although pertinent, exhibits a less robust correlation with other variables. These findings offer significant insights for pension institutions and indicate the necessity for tailored tactics for various social groupings.

Data That Saves Lives? Statistical Capacity as a Pillar of Health Governance: Evidence from 99 Countries

Wed, 31 Dec 2025 00:00:00 GMT

This study investigates whether stronger national statistical systems are associated with changes in recorded health outcomes, focusing primarily on maternal mortality and secondarily on child mortality across 99 countries from 2013 to 2023. Motivated by the heightened awareness of data gaps during the COVID-19 pandemic, the analys s evaluates whether statistical infrastructure—measured through the World Bank's Statistical Performance Indicator (SPI_ALL)—is associated with changes in recorded preventable mortality, consistent with differences in reporting completeness and governance capacity. Using dynamic panel estimators, specifically the two-step System GMM estimator within the Arellano–Bond / Blundell–Bond framework, we control for macroeconomic, institutional, and demographic factors. Results show that higher statistical capacity is significantly associated with higher recorded maternal mortality across both low- and high-income countries. In contrast, its effect on child mortality is statistically significant only in high-income settings. This asymmetry highlights the role of institutional capacity in mediating the effectiveness of data systems, particularly governments’ ability to translate information into actionable governance responses. The findings suggest that statistical capacity functions not merely as a monitoring tool but as a foundational component of public health governance. By linking national data systems to improvements in the measurement, coverage, and usability of mortality statistics, this study supports WHO's strategy for interoperable health information systems and reinforces global priorities under the 2030 Sustainable Development Goals (SDGs).

The Evolution of Medicine Prices in Romania – An Analysis on the Consumer Price Index as Against December Using Jdemetra+ 2.2.4

Wed, 31 Dec 2025 00:00:00 GMT

The pricing of medicines has become a critical issue across the world, which has consequences on both healthcare accessibility and economic stability. In this regard, policymakers, healthcare providers, and the general public should understand the price change dynamics to ensure appropriate prices for the medicines and access to them by all. The focus in this paper is on the Consumer Price Index for medicines in Romania, for the January 2019 to December 2022 timeframe. In this research, JDemetra+ 2.2.4 has been used for seasonal adjustment in assessing the impacts of economic policy. Results show important seasonal variations which significantly influence the role that CPI plays as an indicator of inflation and its possible use for policy making. These findings highlight on the necessity of using seasonal adjustment methodologies for capturing correct price trends in the conditions of changing economic dynamics.

Uncovering Patterns of Informality and Skills in the EU Through Cluster Analysis

Wed, 31 Dec 2025 00:00:00 GMT

The informal economy remains a persistent structural challenge across European countries, closely linked to governance quality, socio-economic development and skill disparities. This study examines how informality relates to skills by combining four alternative estimates of the informal economy with a wide set of educational, digital and institutional indicators. Using Elastic Net regularisation, we identify the most robust determinants of economic performance and subsequently apply k-means clustering to classify 26 European countries according to their profiles of informality and skills. Tertiary education, regulatory quality, digitalisation and life expectancy emerge as recurrent structural factors. The results consistently reveal two clusters: one comprising high-capacity, high-skill, low-informality countries in Western and Northern Europe, and another including lower-skill, higher-informality countries in Southern and Eastern Europe. The study contributes new evidence to a relatively limited European literature on informality and skills, highlighting the need for more granular regional analyses to capture internal heterogeneity and better inform targeted policy interventions.

Zipf's Law and the Evolution of Musical Lyrics: A Comparative Analysis of the 1960s, 1990s and the Present

Wed, 31 Dec 2025 00:00:00 GMT

Song lyrics offer valuable insights into the social and cultural contexts of each period, reflecting the themes and concerns of the time. This study applies Zipf's law to analyse the evolution of lyrical complexity using data from the Billboard Year-End Hot 100 singles during three distinct periods: 1959–1969, 1991–1999, and 2016–2024. To assess repetitiveness trends, the analysis employs the Zipf coefficient (α̂) and the coefficient of determination (R²) derived from log-log models. The findings reveal lower repetitiveness in both the 1960s and the contemporary period, according to the value of the coefficient α, with higher degree observed in the 1990s. Contrary to the initial hypothesis, the results suggest that the digital age has not contributed to the simplification of lyrics. This research underscores the influence of technology on music production and opens avenues for further exploration into its impact on other forms of art and culture.

The Price Elasticity of Electricity Demand in Romania

Wed, 31 Dec 2025 00:00:00 GMT

This paper examines the causal relationships in electricity consumption patterns and price using Granger causality analysis. Our methodology implements vector autoregression models while accounting for electricity's nature as an essential resource with low price elasticity. Through Seasonal Autoregressive Integrated Moving Average (SARIMA) modelling and stationarity assessments, we establish that electricity consumption demonstrates significant inelasticity to price fluctuations, with price series integrated of order I(1) and consumption series of order I(2). Our findings reveal a unidirectional Granger causality from consumption to price, while no significant causality exists from price to consumption. These results suggest that consumption behaviour is primarily influenced by structural and seasonal factors rather than price signals. Furthermore the findings of this article contribute to understanding the dynamics of the electricity market and opens a door to understanding the policy formulation of the governmentsduring the transition to renewable energy.

Determinants of Employability in the EU: The Role of majority Attitudes and Perceptions Toward Roma Communities

Wed, 31 Dec 2025 00:00:00 GMT

This study explores the factors affecting employability among the general population within the European Union by analysing how socio-demographic traits, household situations, and mainstream attitudes towards Roma communities influence labour market outcomes. Using microdata from Eurobarometer 91.4 (2019), the analysis incorporates descriptive statistics, non-parametric tests, and binary logistic regression, complemented by probit regression and linear discriminant analysis as robustness checks.Findings show that older individuals, married respondents, and those in larger households are more likely to be employed, while financial hardships, rural living, and early career changes decrease these chances. Gender disparities still favour men. Attitudinal factors also play a significant role: limited contact with Roma people, low awareness of Roma in schools, and uncertainty about discrimination are linked to poorer employment prospects, whereas integrated labour market participants exhibit higher social distance regarding minority political representation, suggesting a complex socio-political cleavage. The study concludes that labour-market integration depends not only on structural elements but also on social visibility and intergroup familiarity, highlighting the need for policies that address both economic barriers and societal perceptions within the majority population.

Overview of disruptive technologies. A bibliometric approach

Fri, 01 Aug 2025 00:00:00 GMT

Disruptive technologies have integrated into all aspects of our modern lives, rapidly transforming industries and economies on a global scale. Although this phenomenon has attracted considerable attention from researchers and specialists across various fields, there is still an acute need for a detailed and comprehensive analysis. These technologies are both expansive, having a large-scale impact, and specific, acting differently depending on the sector in which they are implemented. This research aims to conduct a rigorous bibliographic analysis, using the Bibliometrix R package and VOSviewer, to understand the fundamental concepts underlying these innovative technologies and their disruptive nature. The study will deeply investigate how disruptive technologies modify traditional economic structures, causing essential changes in the way businesses operate and how economies develop. Additionally, the key findings include the identification of major conceptual clusters, such as innovation management, digital transformation, and sustainability, and the emergence of technologies like blockchain, cloud computing, and additive manufacturing. Trend analysis also highlights a shift in focus over time, with formerly central topics like the internet declining in prominence.

Unraveling the link between E-Government and sustainable development indicators in Europe with the PLS path model

Fri, 30 May 2025 00:00:00 GMT

As we see around us, technology is part of our lives and will gain more and more ground, leading to technology in all areas of life. In this context, digitalization plays a crucial role for the public sector and the citizens, especially young people who are the most connected generation in online communication. Very often, when discussing state institutions, it comes to light the need for more efficient changes since the interaction with public administration is very complex. The purpose of this study is to make a significant contribution to the digitalization literature by exploring how sustainable development indicators influence governance digitalization performance. The conceptual framework of this study combines the 2022 United Nations E-Government predictors with several 2022 World Bank world development indicators. The partial least squares path modeling method was used to explore 2022 secondary data for thirty-six European countries. The findings of this study reveal that technology usage, trust in the public sector, decent work and economic growth and education duration have a significant impact on the digital government performance. Therefore, these insights offer valuable practical implications for optimizing E-Government functioning in the European countries.

Abstract versus Full Paper: A quantitative approach

Fri, 30 May 2025 00:00:00 GMT

The purpose of this article is to observe the differences between the words from selected papers and the terms from summaries or introductory chapters, that correspond with the selected papers, using quantitative methods. To test the above hypothesis, we took the terms from abstracts and introductory chapters, of three different type of papers (a scientific article, a literature book and a methodological study), and we compared them with the most frequent words from the selected papers. Furthermore, we also applied different methods from text mining, such as calculating the phi coefficient or checking compliance with a specific empirical law (Zipf, Heaps), to spot the existent dissimilarities. Following this analysis, it can be observed, from a quantitative point of view, that certain frequent words from the complete writings are found in their introductory components, present a fairly high frequency and outline the main ideas from the respective texts. This paper was co-financed by The Bucharest University of Economic Studies during the PhD program.

Exploring the Efficacy and Bias of Sentiment Analysis Tools in Customer Reviews

Fri, 30 May 2025 00:00:00 GMT

Sentiment analysis is an important technique for understanding public sentiment in customer feedback. This study aims to compare the effectiveness and reliability of BERT and VADER in analysing customer reviews from an e-commerce platform. We analyzed mean sentiment scores and standard deviations for each tool, using Spearman’s Rho to assess correlations between VADER and BERT results due to non-normal score distributions. Our findings reveal that VADER predominantly categorizes reviews as Neutral, with limited Positive and Negative classifications, while BERT shows a bias towards Negative sentiments with some Positive and no Neutral classifications. A weak positive correlation indicates limited agreement between the two tools. Additionally, neither tool showed a strong correlation with actual review ratings, highlighting challenges in accurately capturing nuanced sentiments. The study discusses potential biases inherent in each tool’s methodology – VADER’s lexicon-based approach may oversimplify sentiments, and BERT’s deep learning architecture may introduce a negative bias. These findings suggest that combining VADER’s efficiency with BERT’s contextual sensitivity – for instance, through hybrid models – can mitigate individual tool limitations and offer a more robust sentiment analysis framework for e-commerce businesses. By integrating these insights, businesses can optimize customer feedback systems, refine marketing strategies, and improve product development processes

Resilience Analysis of the European Insurance Industry

Sun, 23 Feb 2025 00:00:00 GMT

The insurance market does not generate only a risk dispersion tool, but constitutes a complex system that contributes to the efficient and effective functioning of the economy. In order to guarantee the fundamental role of this market in sustainable economic growth, insurers must prove resilience in the face of risk and uncertainty generated by the complexity of the system of interdependencies in which they operate. The purpose of the article is to analyze the resilience of the European insurance market, its adaptability to changes and the recovery of delays caused by recent health (especially the COVID-19 pandemic). The data used include variables that define the European insurance system (such as gross written premiums, gross claims incurred, total technical expenses, total assets, total liabilities and number of entities) from EIOPA’s annual statistics from 2017 to2023. Analysis tools include cluster analysis to evaluate the dissimilarity and similarity of European Union countries. The resilience of the insurance industry to shocks was also analyzed from the perspective of financial performance by using a clusterization analysis in order to group countries based on Euclidean distance. The results indicate a grouping of countries in two clusters whose composition remained relatively stable to shocks.

Understanding Young Informal Workers in the European Union: Pre-Pandemic Labour Market Dynamics

Sun, 23 Feb 2025 00:00:00 GMT

Young people are one of the most important groups in the labour market, but also one of the most vulnerable. They do not find it easy to enter the labour market, which may have a negative impact on the growth of undeclared work among this group. The main objective of this study is to establish a baseline understanding of the situation of young informal workers prior to the pandemic by developing a comprehensive profile of these workers at the EU27 level. This profile is based on statistical analysis of undeclared work data from Eurobarometer 92.1/2019. The results of this study include: the share of young informal workers in the EU27, the types of activities they perform, the sectors of the labour market in which they perform these activities and the reasons why they perform them. By creating this profile, this study aims to contribute to future research that will provide a clearer understanding of how the Covid-19 pandemic has led to labour market changes in youth informality.

Gender Gaps in Undergraduate Mathematics Performance in Romania: An Analysis Using General Linear Models

Sun, 23 Feb 2025 00:00:00 GMT

This paper investigates gender gaps in mathematics performance among Romanian Baccalaureate students across four educational profiles: Mathematics-Informatics, Natural Sciences, Technological, and Pedagogical, between 2017 and 2023. Using descriptive statistics, and statistical analyses such as t-tests and General Linear Models, the study reveals consistent gender disparities in mathematics performance, with female students generally outperforming male students. The results highlight significant variations in performance across the different educational profiles, pointing to a need for profile-specific educational interventions. The spatial analysis underscores the need for targeted strategies that address not only gender gaps, but also regional inequalities. The findings suggest that Romania’s policy of equal access to education may not fully address the differences in educational outcomes, thereby calling for tailored interventions to bridge the performance gaps across gender, profile, and region.

Exploring the Labor Market Implications of the Pandemic for Women

Sun, 23 Feb 2025 00:00:00 GMT

The labour market has been profoundly impacted by the pandemic. Efforts to preserve employment during lockdowns have led to a significant shift toward remote work across various sectors, intensifying challenges in balancing work and family responsibilities. Our paper investigates how women were affected by the pandemic in terms of work and family. We rely on data from the Eurobarometer ‘Women in times of COVID-19’ to assess the experiences of women aged 15 and above in the 27 Member States of the European Union. A logistic regression-based approach is employed to evaluate which women were most impacted by the pandemic in terms of labour market participation and income reductions. The results revealed that women in rural areas or those being self-employed were more likely to be vulnerable. Furthermore, women caring for children under 15 encountered substantial difficulties in the labour market. Income reduction and decreased labour market participation were also challenging for women who experienced a negative impact on their mental health due to government measures during the pandemic or those who became more financially dependent on partners or family members.

Analysis of Vulnerable Migration in European Union Countries Using a Gravity Spatial Model: The Syrian Case

Tue, 24 Dec 2024 00:00:00 GMT

Migration represents a complex phenomenon that, when associated with vulnerability, becomes a theme of particular importance in the current global context.The refugees’ crisis had a direct influence on permanent migrants. The COVID-19 pandemic had an impact on short-term migration, increasing the quantity of short-term migrants while decreasing the number of long-term displacement. In this context, the data presented in our study are from 2019, a year in which the mobility constraints resulting from the pandemics have not yet had an impact on migratory flows.All European Union countries were covered, except Cyprus, where data was unavailable. Our study uses a gravity model of migration Syrian migrant. Through our analysis, we aimed to capture the determinants of migration flows from Syria to EU countries. It also aims to capture a range of spatial effects that may have an impact on the factors influencing the migration flows of Syrian migrants into the European Union countries. The results showed significant influences on Syrian migration flow to European Union countries, distance being a decisive factor in the decision to migrate.