Modelling the quarterly GDP - Role of economic and surveys indicators

The quarterly GDP forecasting models developed in this material aim to estimate the Albanian GDP trends in the short term. . Delays until the publication of the official quarterly GDP data make indispensable the preliminary estimation of this indicator. The modelling strategy of the quarterly GDP consists in building a set of different models for its estimation. They consist on ARIMA models with seasonal components and indicator models, similar to bridge models. This paper presents a first attempt to model the GDP using a multiequations system which accounts for the sectoral interactions. This model can not be used for forecasting purposes because of short time series. The estimates were made for total and for disaggregated sectoral GDP for the period: Q1:2003 – Q1:2009. The models exploit information from economic variables, financial variables and confidence surveys indicators, held by the Bank of Albania. The bridge models estimates show that the past developments of economic and financial variables explain the GDP changes while the survey variables lead them. The above mentioned behaviour of the explanatory variables supports the forecasting process of the quarterly GDP. Thus the policy makers in the Bank of Albania are provided with a timely estimation (nowcast) of the economic activity tendency for the reference quarter and for the coming two quarters. In general, estimations from the developed models are promising. It is suggested that the “best” forecast will be considered the average of the forecasts from all the proposed models. The off-sample forecast performance, “will decide” the model with the best qualities in order to predict the quarterly GDP.