Artificial Intelligence in economic decision making: how to assure a trust?

In this study, we analyzed in situ data from the years 2018–2022 encompassing entire forest plantations in Poland. Based on data regarding stand density and the occurrence of fungal, water-related, climate-related, fire, and insect factors that may intensify with climate changes, we determined the correlation between their occurrence and the decline in wood increments for six tree species: pine, birch, oak, spruce, beech, and alder. Subsequently, we identified age intervals in which the species–factor interaction exhibited statistically significant effects. Next, we developed neural network models for short-term wood increment predictions. Utilizing these models, we estimated a reduction in wood supply harvested in accordance with the plans for the years 2023–2025 assuming a tenfold greater intensity of factors than in 2022. Findings indicate: birch: water-related factors may reduce wood production by 0.1%–0.2%. This aligns with previous research linking drought to birch wood decline, highlighting its sensitivity to water-related issues. Oak: fungal and insect factors could decrease wood production by up to 0.1%. Prior studies emphasize the significant influence of fungal diseases on oak health and regeneration, as well as the impact of insect infestation on wood production. Alder: water-related factors may lead to a slight reduction in wood production, approximately 0.02%. The impact is significant within specific age ranges, indicating potential effects on harvesting. Pine: water- and climate-related factors may result in up to a 0.05% reduction in wood production. Pine, a key forest-forming species in Poland, is notably sensitive to these factors, especially as it nears harvesting age. Spruce: insects, fungi, and climate-related factors could lead to a reduction in wood production of up to 0.2%–0.3%. Analyses demonstrate sensitivity, resulting in a noticeable growth differential compared to the typical rate. Short-term predictions based on neural networks were developed, acknowledging their suitability for short-term forecasts due to uncertainties regarding long-term factor impacts. Additionally, our study discussed modeling wood increments in divisions well below the harvesting time, emphasizing that the influence of current and 2023–2025 factors on wood increments and supply may only manifest several decades from now. These results imply important indications for the economic and financial performance of the wood industry.

Liquid fuels obtained in refining crude oil are one of the most important energies in economic activity. The domestic wholesale market for liquid fuels is of decisive importance for price formation in the national economy. The noncompetitive behavior of the market players at this level of the distribution chain can significantly affect all downstream price levels and the producer–consumer surplus balance. Therefore, the competitiveness of this market should be screened and assessed regularly, especially when significant external factors change. This article attempts to evaluate the impact of structural changes on the global market of crude oil and energy products after the outbreak of the COVID-19 pandemic on the competitiveness of the wholesale fuel market in Poland. Using asymmetry of the reaction of product prices to changes in the prices of inputs as a marker of noncompetitive behavior and the NARDL model as a test specification, the price paths of market players before and after the occurrence of structural changes in the inputs’ processes were examined. Significant changes in the competitive behavior of players were revealed after the occurrence of structural changes at the beginning of the pandemic period in the year 2020. These changes may indicate enhanced competition and mitigation of potential market power abuse.

The detection and deterrence of collusive agreements among firms, such as price-fixing cartels, remain pivotal in maintaining market competition. This study investigates the application of machine learning methodologies in the behavioral screening process for detecting collusion, with a specific focus on parallel pricing behaviors in the wholesale fuel market. By employing unsupervised learning techniques, this research aims to identify patterns indicative of collusion—referred to as collusion markers—within time series data. This paper outlines a comprehensive screening research plan based on the CRISP-DM model, detailing phases from business understanding to monitoring. It emphasizes the significance of machine learning methods, including distance measures, motifs, discords, and semantic segmentation, in uncovering these patterns. A case study of the Polish wholesale fuel market illustrates the practical application of these techniques, demonstrating how anomalies and regime changes in price behavior can signal potential collusion. The findings suggest that unsupervised machine learning methods offer a robust alternative to traditional statistical and econometric tools, particularly due to their ability to process large and complex datasets without predefined models. This research concludes that these methods can significantly enhance the detection of collusive behaviors, providing valuable insights for antitrust authorities.

Motivation: Import parity pricing, potentially provoking parallel pricing, and asymmetric price transmission are both considered as the factors that can distort competitiveness and foster collusive behavior on the fuel markets. Aim: The study encompass an analysis of usual ‘rocket and feathers’ reactions in response to common drivers of the gasoline wholesale price and comparative analysis of individual players’ pricing paths in order to examine competitive performance of the market. To account for an implied IPP pricing mechanism, a set of common wholesale price determinants was expanded. Results: We revealed a significant short-run asymmetries in a transmission of all downstream price determinants and shown that the foreign exchange rate is a main driver underlying a positive asymmetry in the wholesale prices’ paths. We compared dynamic price adjustment paths for the major players and revealed the common patterns. That suggest a strong possibility of parallel pricing.

Anticompetitive horizontal behaviors are very common and harmful for economy as a whole and consumer welfare especially. Such behaviors in most cases mean collusion. Pro-active anticompetitive policy we want to realize implies monitoring of huge set of markets and industries. As preliminary selection is possible on a basis of number of factors that enhanced collusion, there could be still big set of industries left. There is a necessity then for methods of collusion’s detection which would be quick, resource — saving and understandable for economists, experts and lawyers without quantitative experience. We described here fairly simple but meaningful, graphical tools which could be used in process of screening for collusive behavior in the markets. We underlined that ability of graphical inspection depends on statistical/econometric tool of choice for particular screen, so we see a need for research and popularization of tools with that property.