RACE: Revista de Administração, Contabilidade e Economia (Aug 2020)
Factors determining the forecast errors of market analysts for fiscal variables in Brazil
Abstract
The objective of this study is to investigate determinant factors the forecast errors of market analysts for Brazilian fiscal variables. The data for conducting the research was obtained in the Prisma Fiscal, the Ministry of Economy's system of collecting and disclosure of market expectations for fiscal variables. The data collected refer to collection, net revenue and total expenditure, in the period from November 2015 to December 2018. The Mean Absolute Percentage Error (MAPE) and z were used to measure the quality of the market analysts' forecast. The use of the z value as a measure of the forecast error is one of the contributions of this research. Among the results obtained, the hypothesis that the temporal horizon interferes in the quality of the forecast was not rejected for horizons of one and two years; the dispersion of forecasts did not show a substantial change; and the optimistic bias hypothesis was not confirmed. It can be concluded that for this sample the temporality is a determinant factor of the forecast error of the market analysts for fiscal variables. The research contributes to the discussion about forecasting error in the areas of Public Financial Management and Public Accounting.
Keywords