We develop a methodology for using intra-annual data to forecast annual budget deficits. Our approach aims at improving the accuracy of the deficit forecasts, a relevant issue to policy makers in the Eurozone and at proposing a replicable methodology using at best public quantitative information on budgetary data. Using French data on government (State) revenues and expenditures, we estimate intra-annual monthly ARIMA models for all the items of the central government revenues and expenditures. Next, applying temporal aggregation techniques, we infer parameters of the annual models from the estimated parameters of the intra-annual models. These parameters incorporate all the intra-annual information. Finally, we do one period ahead predictions. We are able to update the annual deficit forecast as soon as new monthly data are available. This allows us to detect possible slippages in central government finances.
Moulin, L., Salto, M., Silvestrini, A., & Veredas, D. (2004). Using intra annual information to forecast the annual state deficits. The case of France (CORE Discussion Papers 2004/48). https://hdl.handle.net/2078.5/128782