摘要：This paper examines the power of business and consumer survey indicators in predicting retail activity in Croatia. A brief overview of business and consumer surveys is presented, followed by the conceptual logic behind their use in predicting related macroeconomic variables. Detailed elaboration of methodology used is presented next. The main model used for the analysis is a bivariate vector autoregression VAR together with causality tests, impulse response function analysis and variance decomposition. Results have revealed that a statistically significant relationship exists between consumer sentiment indicator and retail activity variable. To an immense surprise, that relationship turned out to be negative. Possible explanations for such an unexpected result are suggested.
关键词：Business and consumer survey; vector autoregression model; retail activity
