Forecasting growth in eastern Europe and central Asia

Growth forecasts are at the heart of investment decisions and yet in emerging and developing markets they often receive limited support from quantitative models. This paper describes the main quantitative modelling techniques that have informed EBRD growth projections since 2009. The paper describes how bridge models use high-frequency data and the generalised dynamic factor model uses a vast body of cross-country data to nowcast growth. In addition, a Bayesian vector-autoregressive model is used to forecast growth under various scenarios. The paper includes a comparison between the forecasting performance of the model and that of economists, the latter working both at the EBRD and at other institutions.