Nowcasting quarterly GDP in Albania

The quarterly Gross Domestic Product in Albania is published with a lag of 12 weeks (three months) from the reference quarter. In between, it is important for policy makers to have an evaluation of the aggregate demand and domestic output, to be able to take wellinformed decisions and implement economic policy effectively. For this purpose, this study tests the predicting power of different economic indicators that are available before the quarterly GDP publication. The immediate-term forecasting, a.k.a. nowcasting, is obtained with ADL (p, q) equations, and estimated with OLS. Main results reveal that ‘auto sales and repairs’- a hard indicator -, has the highest predicting power among all tested variables, followed by qualitative (survey) indicators like ‘the industrial sector’s evaluation of current demand’, ‘the industrial sector’s evaluation of current production’, ‘the industrial sector’s expectations on next quarter’s production’, ‘the construction sector’s evaluation of current production’, and ‘the services sector’s evaluation of current demand’. These qualitative indicators are disaggregated indices of the Bank of Albania’s Consumer and Business Surveys. The final nowcasting model includes ‘auto sales and repairs’ and ‘the construction sector’s evaluation of current production’. Selected candidate models outperform naïve and ARIMA models. Keywords: nowcasting, qualitative indicators, Gross Domestic Product, Albania JEL classification: C22, C53, E17