Пространственное моделирование электоральных предпочтений в Российской Федерации

Пространственная экономика. 2020. Т. 16. № 2. С. 70–100.

Podkolzina E., Demidova O., Kuletskaya L.

The main objective of this work is to assess the influence of individuals living in neighboring territorial areas on each other in decision-making on the example of presidential election in Russia in 2018 using data on 2718 territorial election commissions (TECs). Local and global indicators of spatial autocorrelation (Moran, Geary, GetisOrd indices) calculated by the authors provide empirical evidence of global positive autocorrelation (i.e. in the country as a whole voters in each TEC vote similar to their neighbors). We identify TECs that can be included in local clusters (where voters vote similar) or in local outliers (surrounded by such TECs where voters vote opposite. Using the example of Tatarstan, the region where both local cluster and outlier TECs were most common we analyzed which economic indicators together with spatial ones influence the support of the main and opposition candidates. It was shown that the willingness to vote for the main candidate is explained by the increase in salaries in the area, but at the same time the indicators of economic activity in that area and the potential mobility of citizens have a negative impact on the support of the main candidate. Salary changes have no effect on votes in favour of opposition candidates, while other indicators show an inverse correlation. We have also shown that spatial effect models are preferable to OLS models for analyzing voting results.

Keywords: пространственная автокорреляция, электоральные предпочтения, spatial autocorrelation, electoral preferences, global and local indices of spatial autocorrelation, глобальные и локальные индексы пространственной автокорреляции

Publication based on the results of: ­­­ Analysis of incentives and preferences of economic agents in higher education, in the banking services and in public procurement (2020)