Resumen de Forecasting Corporate Income Tax Revenues: the incidence of leading and soft indicators in a time series framework

Camino González Vasco, Raquel Pajares Rojo

• This paper contributes to the empirical literature on the development and estimation of a forecasting model for Corporate Income Tax (CIT) revenue focusing on the study of macroeconomic indicators. The proposed study uses a set of predictors comprising Central Bank Balance Sheet Data and soft indicators, along with some macroeconomic variables related to domestic demand and the foreign sector, which amounts to a total of 43 initial quarterly variables. To address the problem of multicollinearity, we use a blend of statistical methods such as Principal Components Analysis and ARX (Auto-Regressive with eXogenous input) models. Results of the backtesting exercises show that this combination successfully predicts the evolution of the CIT revenue and illustrate its usefulness as a tool for short-term CIT revenue forecasting.