Estimation and forecast of the current account balance in Azerbaijan

The purpose of the research is to identify the main determinants of the components of the current account balance (CAB) in Azerbaijan and build a forecasting model for CAB at the disaggregated level. Individual sub-components of CAB are estimated using a wide range of different macroeconomic variables. For estimation purposes quarterly data for the period 2003Q1-2020Q4 is used and then the forecasts for the next year are obtained. The econometric technique employed in the research is autoregressive distributed lag (ARDL) model. 3 forecast scenarios are generated under differing assumptions on oil price, on government spending, proxied by M2 monetary aggregates, as well as on growth projections. According to the results of the estimation, the variables showing the most significance in explaining the changes in different CAB subcomponents are found to be oil prices, oil production, government spending, non-oil REER, income of population and the lags of the respective dependent variables. Overall, the model allows us to identify main determinants of CAB components and forecast CAB. Especially, the accuracy of the forecast of trade balance is high. However, for a few components the forecasting accuracy appears to be low, as they suffer from high forecast errors.