Measurement and modelling of business cycles using linear and nonlinear methods

Abstract

This paper focuses on business cycle component analysis and modelling in small economy countries. Lithuanian economy is one of such example with specific properties. The list of major leading indicators business cycles was studied in this paper in the case of Lithuania. The results suggest that economical methods aren’t suitable for modelling of main indicators business cycles in such cases. Standard univariate methods are preferable. However various time series business cycles may be nonlinear, therefore linear ARIMA and nonlinear SETAR methods were used and availability to capture features of the real time series business cycle with complex nature was tested.