Testing for Business Cycle Asymmetries : The Case of Croatia
We analyze asymmetries of the business cycle in the Croatian economy, using various annualized growth rate variables for a period of eighteen years (1992-2010), de-trended by Hodrick-Prescott filter, and following the Harding and Pagan procedure in determination of its turning points. The time series in which we are looking for asymmetries over the Croatian business cycle are monthly data sets: gross domestic product (GDP); industry production index (IND); construction production index (CONST); nominal total import in Euro (IMP); nominal total export in Euro (EXP); tourism arrivals (ARR); tourism overnight stays (ON); retail consumer price index (RPCI); nominal wage (WAGE); productivity in industry index (PROD); nominal narrow money (M1); real exchange rate (REAL); real NB discount rate (INTR); unemployment rate (UR); vacancy rate (VR). In testing the issue of asymmetries we used and the procedure consists of the application of the formal Sichel's tests, which allows for two types of asymmetry in mean to be tested: steepness and deepness. Steepness occurs when contractions are steeper than expansions, whilst deepness is found when troughs are deeper than peak heights. We explore whether the evidence of asymmetries can be associated to an asymmetric behavior over the business cycle by augmented autoregressive models, too