Cyclic dynamic patterns of Russian macroeconomic indicators found by spectral analysis

Abstract

The paper proposes a contemporary interdisciplinary method to identify consistent patterns within cyclical dynamics of GDP and its macroeconomics determinants in the Russian Federation. This method may contribute to better recognition of the stages of economic cycle and of potential early predicators to recessions and crises. We first identify the trend component of Russian GDP and then apply the spectral data analysis to its cyclical component which reveals its multi-frequency, and non-linear vibrations. These vibrations are then further investigated by transforming time series data on GDP and its determinants into a frequency spectrum series via Fourier transform techniques. Wavelength scanning of selected macroeconomic indicators shows the basic economic cycle of real GDP with duration time of approx. 3.13 years. Other procyclical indicators nevertheless discover asynchronous behavior towards GDP due to the relative autonomy of the sectors standing behind these indicators. Their autonomy lies behind differences in reaction forces (shifts) and periods (lags) to both internal and external shocks. We estimate differentials between the dynamics of GDP and its determinants by evaluating phase deviations of their pairwise harmonic components, mutual pairwise phase shifts, and by comparison of their pairwise cross-spectrum. The one of output is the quantification of time lags between GDP and key macroeconomic indicators of individual economic sectors. This result reveals the complexity of GDP dynamics that sends an aliased rather than a unit signal to economic agents. Our decomposition of this signal into signals from key economic sectors and quantification of phase discrepancies between sectoral signals may contribute to findings in early crisis predicators. We also estimate the depth and velocity of shocks penetrations into both economy as a whole and its particular sectors.