Title: Enhancing economic forecasts through Hilbert space projection methods Authors:  Tsvetomira Tsenova - Experian Bulgaria (Bulgaria) [presenting]
Abstract: Improving forecast accuracy of economic and financial variables is essential in central banking and financial institutions, as conditioning decisions on economic developments has become obligatory. Forecast reports contain the entire probability distribution of the target variables, i.e., point, uncertainty, risk, derived probabilistic scenarios, at near, medium, and long-term horizons. However, traditional forecasting methods are known to converge to their equilibrium much too quickly. While having sensible equilibrium properties is desirable in banking and finance, the decision-makers are deprived of the opportunity to design an optimal reaction to the shorter-term predictable volatility. The aim is to enhance the applied forecasting process by exploiting the Hilbert Space Projection methods. Advantages and disadvantages of various forecasting methods are discussed from a practical perspective, and optimal ways to combine them.