Keywords

nowcasting GDP, nowcasting, factor models, Kalman filter

This paper tests of the extent to which available monthly economic indicators help in nowcasting GDP. For this purpose, a factor model is proposed on data relevant for the movement of the domestic GDP (monthly indicator of real economic activity - MRGA) and its results are compared recursively with models for nowcasting from recent related literature and with simple benchmark models. The evaluation of the results of the model indicates that factor models based on the dynamics of a broad group of variables produce better nowcasts than the benchmark models used. Furthermore, different specifications of factor models have similar performance. Another important finding of the analysis is that it is worthwhile to combine the information available in individual models. In addition to nowcasting, monthly series of GDP growth rates for Croatia based on the movement of a large number of available monthly indicators are also constructed in the paper.