Spatial Analysis of the Relationship Between Income and Consumption of the Population Based on Panel Data

Abstract

The article presents results of a study on influence of population dynamics, regional characteristics and the type structure of income on consumption. The ability to investigate spatial dependencies and territorial effects over time was made possible by autoregression spatial models built on panel data. The article describes features of such models, sequence of calculations, and also presents modified tests to justify the choice of the model specification.Calculations were made using data from 83 constituent entities of the Russian Federation (cross-sectional observations) for 2010–2019 (10 time periods). The analysis showed that both population income and retail turnover, which largely determine the level and structure of population consumption, have spatial dependencies. The built spatial error model with fixed effects showed a positive influence on population consumption in the neighboring territories. The model also confirmed previously identified relationships: the positive impact of average per capita income and the negative impact of the Gini index on consumption. The built model with fixed effects allowed to isolate the individual effects of the territories, visualized using cartogram. On the basis of these assessments, several groups of territories with common properties and characteristics have been identified.Unlike previously built models, the authors’ spatial error autoregression model, built on panel data, took into account both the geographical heterogeneity and spatial dependence of average per capita income and retail turnover, expanding the existing understanding of the relationship between consumption and income. This, in turn, enables management decisions that take into account previously undetected features and enhance their validity.