Experimental Composite Indicators of Cyclical Response in Surveys of Consumer Expectations of the Population

Abstract

The article proposes a theoretical rationale and procedure for constructing new composite indicators that combine the results of quarterly consumer expectations surveys conducted by the Federal State Statistics Service (Rosstat). The information capabilities of surveys for the period 2005–2021 were studied to find the best alternative to the traditional consumer confidence index in the context of early estimates of the growth in household final consumption. The experimental Consumer Behavior Index reflects the «common profile» in the changes in the key parameters of consumer activity. The index is based on time series of primary survey indicators with high cross-correlation coefficients and a confirmed statistically significant causal relationship with a statistical referent – the growth rate of the final consumption expenditure of households. The new Consumer Anxiety Index with a counter-directional profile assesses the scale and tendencies of household response to market shocks based on short-term gaps between the indicator components and their long-term average.In accordance with the proposed procedure for statistical testing of the time series, a cross-correlation analysis was carried out. Granger causality was studied, principal component analyses applied, unobservable cyclical components were identified in the dynamics of the Consumer Behavior Index and the referent using Hodrick – Prescott, Christian – Fitzgerald and Kalman statistical filters, and phase-by-phase movement of these components is demonstrated. A statistically significant cyclical correspondence and a close relationship of leading and coinciding nature between the composite survey-based index and the growth rate of the final consumption expenditure of households were confirmed.