VARX model using leading cycles

Abstract

Composite leading indicators serve to the purpose of early identification of the stage of business cycle and its turning points. The indicators are constructed on the basis of statistical analysis for predicting the phase of business cycle via the VARX model approach. The model of Slovakia is composed of three variables: gross domestic product as reference series (endogenous variable), composite leading indicator for the domestic economy (endogenous variable) and composite leading indicator of external environment (exogenous variable). The quality of the model is verified using statistical tests of stationarity and causality of variables and tests of residuals. The VARX model exhibits some ability to adapt to changes of the business cycle, but the quantitative projections are rather unable to predict accurately the turning points.