Hysteresis and Long Memory in Cash Income Dynamics of Russian Regions

Бесплатный доступ

The persistence of regional disparities is a key challenge for Russia’s economic development. The concept of path dependence suggests that historical shocks can have long-lasting, and sometimes irreversible, effects on regional development trends. However, there is a lack of empirical studies systematically assessing the degree of income persistence across all Russian regions using current data and modern time series analysis techniques. This study aims to fill this gap by investigating the presence and nature of persistence (hysteresis and long memory) in the dynamics of real per capita income in Russian regions for the period 2013–2023. To achieve this goal, unit root tests (ADF, KPSS) on the original series were used; MSTL decomposition was used to account for nonlinear trends; unit root tests on the detrended residuals were used; tests for long memory (GPH, ARFIMA estimation) and autocorrelation function (ACF) analysis of the residuals were employed. The analysis results indicate a high degree of persistence in income dynamics for the vast majority of regions. Analysis of the residuals after nonlinear detrending suggests that for a significant portion of regions, income dynamics align with the hysteresis hypothesis (an I(1) process), implying permanent effects of shocks. For another group of regions, trend stationarity appears more likely. Formal long memory tests indicate the presence of stationary long memory (0 < d < 0.5) in the residuals for many regions; however, given the short time series length (11 years) and contradictions with other tests, these findings are interpreted cautiously as further evidence of high persistence. Authors’ contribution. M.V. Dubovik – analysis of literary sources and genesis of key concepts, as well as their conceptualization. S.G. Dmitriev – coding, graphing, analysis of indicators, and interpretation of results.

Еще

Regional income, path dependence, hysteresis, long memory, fractional integration, unit root, time series, Russian regions

Короткий адрес: https://sciup.org/149149355

IDR: 149149355   |   УДК: 330.42   |   DOI: 10.15688/re.volsu.2025.3.2