
Risk is one of the fundamental concepts analyzed in economics and constitutes an inherent element of economic activity. In practice, particular importance is attached to the risk of occupational accidents, which may result in loss of life, injury, deterioration of health, and reduced ability of farmers to achieve their economic objectives. The main objective of this study was to assess changes in the level and structure of occupational accidents in agriculture in Poland between 2014 and 2024, taking into account the type of accident, farm size, and agricultural production profile. The study covered all occupational accidents registered by the Agricultural Social Insurance Fund (KRUS) during the eleven-year study period (2014-2024). The source material was based on a review of the relevant literature and unpublished secondary data obtained from KRUS. The collected data were analyzed using trend analysis. The results confirmed a long-term downward trend in the number of occupational accidents in agriculture. Despite favourable quantitative changes, the structure of accidents by event category remained relatively stable. Falls of persons constituted the dominant category throughout the analyzed period. These findings indicate that preventive measures should focus more strongly on reducing the risk of falls during agricultural work. The results also demonstrate that the structure of occupational accidents varies with farm size and indicate that the smallest farms require particular support in improving occupational safety. The study contributes to filling the research gap regarding long-term trends in occupational accidents in agriculture.
This study evaluates the forecasting performance of Random Forest (RF) and k-Nearest Neighbors (kNN) models for weekly wholesale prices of Peruvian yellow potato (Solanum goniocalyx) in Lima over one- to four-week forecast horizons. The analysis uses 238 weekly observations from 8 January 2021 to 25 July 2025, transformed into a supervised learning dataset incorporating lagged, rolling, and calendar-based features. Model performance is evaluated using a strictly time-ordered data split and benchmarked against naive and seasonal naive forecasting approaches. The results show that RF provides the best overall forecasting performance and greater robustness to recent shifts in price levels, while kNN remains competitive for selected short-term forecast horizons. The forecast price path remains close to recently observed levels, while forecast uncertainty gradually increases as the forecast horizon extends. These findings support the use of RF as the primary forecasting model, with kNN serving as a complementary benchmark for short-term price monitoring, purchase planning, and early-warning applications. The study is primarily empirical and application-oriented rather than methodological and provides evidence on the relative forecasting performance and practical usefulness of these established machine learning approaches in the wholesale market for an important staple food product.
The adoption of digital agricultural technologies is becoming increasingly important for enhancing agricultural productivity, improving resource-use efficiency, and supporting the digital transformation of the agricultural sector. However, empirical evidence on the factors influencing farmers’ adoption of such technologies remains limited in Uzbekistan and other developing agricultural economies. This study examines the factors associated with the adoption of digital agricultural technologies among agricultural producers in Uzbekistan by extending the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2) framework to include Trust and Perceived Risk. Primary data were collected from agricultural producers using a structured questionnaire. A total of 373 responses were obtained, of which 280 valid observations were retained after data screening. The proposed model was assessed using Partial Least Squares Structural Equation Modeling (PLS-SEM). The results indicate that Behavioral Intention is significantly associated with actual technology use, while Facilitating Conditions and Habit also contribute to Use Behavior. Among the factors explaining Behavioral Intention, Perceived Risk was the only statistically significant predictor; however, contrary to theoretical expectations, the relationship was positive rather than negative. The remaining UTAUT2 constructs and Trust did not have statistically significant independent effects on Behavioral Intention. These findings suggest that farmers’ decisions regarding the adoption of digital agricultural technologies in Uzbekistan may be influenced not only by perceived technological benefits but also by uncertainty, perceptions of risk, enabling conditions, and previous experience. The unexpected positive relationship between Perceived Risk and Behavioral Intention should, however, be interpreted with caution and warrants further investigation. The study provides empirical evidence from an underrepresented agricultural context and demonstrates the relevance of incorporating contextual factors such as Trust and Perceived Risk into technology acceptance models. The findings also highlight the importance of strengthening digital infrastructure, providing practical support, and reducing technology-related uncertainty to facilitate the digital transformation of agriculture in Uzbekistan.
This article presents four development scenarios for the agri-food data economy through 2050 and outlines their key social, financial, environmental, and fundamental implications. The article takes a synthetic, scholarly approach, using scenarios developed within the Data4Food2030 project through a multi-stage research process involving public administration, academia, business, and other agri-food system stakeholders. Four alternative visions of the future are presented: S1 – Personalised healthy diets; S2 – Local and transparent value chains; S3 – Strictly regulated high-tech and innovation; and S4 – Environmental benefits through data-driven decisions. Comparisons are made across several dimensions: tools, functions, data types, governance models, public policy instruments, benefits, risks, and public policy challenges. The findings indicate that all scenarios represent viable strategic options, differing primarily in their objectives for data development, data governance models, and the distribution of benefits. Expanding data sharing also entails risks – such as those related to data release, cybersecurity, privacy protection, and the potential exclusion of certain producers. These scenarios can help policymakers and agri-food stakeholders formulate strategies that foster a secure, safe, and inclusive data ecosystem.
This study examined the factors influencing public support for costly greenhouse gas (GHG) mitigation policies in the United Arab Emirates (UAE), an oil-dependent economy in which about 89% of the population are transient residents. An on-line survey collected responses from 1,540 respondents in 2021. The sample was applied to estimate the empirical relationship using ordered logit regression and results allowed calculations of the probability change in the support of costly mitigation programs (the latent dependent variable). Results identify agreement with statements about climate-related ideas as the strongest predictors of policy support. Statements such as: “the world’s climate is changing”, “UAE temperatures will rise”, “the balance of nature is delicate,” and “lower oil prices will reduce household income” are associated with support for costly mitigation. Among socio-demographic characteristics, Filipino and Indian respondents – the two largest expatriate groups – show significantly lower support than other expatriates, while Christian respondents and higher-income respondents signal greater support than respondents indicating other religious affiliations or less income.