The Impact of Industrial Output and Investment on Electricity Consumption in Sverdlovsk Oblast (Russia): Wavelet Analysis of Time Series Accounting for Seasonal Factors

Abstract

This article examines the influence of industrial production and investment on electricity consumption in Sverdlovsk Oblast using multivariate wavelet analysis (MWA) that accounts for seasonal factors. The novelty of the study lies in the application of MWA tools, such as multiple and partial coherence, partial phase difference, and partial wavelet gain coefficient, to identify time-varying causal relationships. The wavelet-based results confirm and extend findings gained through the application of traditional econometric approaches by revealing how these relationships differ across time horizons and frequencies. The multiple coherence analysis shows seasonal cointegration at a frequency corresponding to a four-quarter cycle and the absence of long-term (non-seasonal) cointegration. Partial coherence diagrams suggest that, after controlling for one variable, there is no cointegration between electricity consumption and either industrial output or investment across all frequencies. Partial phase difference analysis reveals the lead-lag structure and phase alignment among the variables, depending on the frequency and time period. Notably, data from 2022–2023, coinciding with the imposition of international sanctions on Russia, offer particularly valuable insights. The study shows that both business cycle theories and related government policies should place greater emphasis on seasonal dynamics. Companies can use the results of wavelet analysis to determine the optimal timing for launching new production capacities.