Cyclical Properties of the Credit and Production in Selected European Countries - a Comparison of Deterministic and Stochastic Cycle Approach

We develop two nonparametric approaches to analyze the empirical properties of economic cycles. The first approach is based on almost periodically correlated time series commonly used in signal processing. Within this framework we depart from standard scheme of analysis that relies on stationarity assumption. The second approach is based on spectral analysis provided the stationarity assumption of cyclical fluctuations. We contribute to the existing literature in both, theoretical and empirical aspects. From theoretical viewpoint we develop methods of formal statistical inference about the main properties of elements of the economic cycle. In the first approach the testing procedure utilizing subsampling approach is proposed. In the second approach the method of analysis of concentration of the spectral mass is developed. Based on the monthly series of the credit aggregate and the industrial production, taken from selected European countries, we discuss the empirical properties of the credit cycle and we compare them with the production cycle. Our empirical findings show substantial diversity of the credit cycle across analysed countries. Also cyclical component in the credit series is identified much stronger than in case of the series of industrial production. Also the production cycles are much more synchronized across countries compared to the credit cycles.