R-ECONOMY

Relevance. Since regional markets are interconnected and influence each other, forecasting price changes for goods and services requires considering both time and location. Economic instability, shifting supply chains, and rising inflation expectations make this research especially relevant. Additionally, the growing need for quick responses to price fluctuations highlights the importance of adopting data-processing methods that enable near real-time analysis. Research Objective. The aim of the study was to analyze the spatial dependence of prices and the presence of brands within the context of cyclically fluctuating demand and supply across different price segments. Data and methods. This study utilized data provided by the online analytics service продажи. рф, which encompasses daily selling (registered) prices for 135 ice-cream brands across 84 Russian regions, spanning the period from January 1, 2021, to December 31, 2023. The analysis examined regional differences in ice cream prices and brand representation, as well as the spatial autocorrelation of prices, particularly in relation to seasonal demand fluctuations. Spatial autocorrelation was assessed using global and local Moran’s I indices, with spatial clusters identified based on these estimates. To explore the cyclicality of spatial autocorrelation, partial autocorrelation functions (ACF and PACF) were used, and the Kruskal-Wallis test was applied. Results. The results of the analysis confirmed the differentiation of regions in terms of ice-cream brand representation, including variations across three price segments: Elite, Standard and Economy. We found a correlation between brand representation and regional population size, but no direct relationship with regional wage levels. Further analysis of individual brand prices and their spatial autocorrelation confirmed the hypothesized presence of spatial autocorrelation and demonstrated an increase in this autocorrelation over the study period. Examination of data cyclicality indicated that time series of average prices and global Moran’s I indices exhibited significant weekly cyclicality, while annual cyclicality was not consistently detected across all analytical methods and only emerged in the analysis of average prices. This suggests that seasonal variations in production and consumption volumes do not necessarily translate into corresponding seasonal fluctuations in prices or their spatial autocorrelation for all product groups. Conclusions. Spatial price dependence is not static; its level and dynamics are significantly influenced by product characteristics, underscoring the necessity of shifting from analyses of aggregate-price indices to analyses of individual product prices. A key methodological contribution of this study is the validation of findings previously observed with more aggregated data (year/month, product group) using highly detailed daily and brand-level data. This approach enhances forecasting accuracy by capturing the full scope of regional variations in consumer behavior.