Forecast Error Decomposition in a Nonlinear Model with Provisional Data

In this paper, we take into consideration some issues related to the use of a nonlinear structural econometric model in the presence of a data revision process. We analyse the consequences on the parameter estimation (consistency is still attainable) and on forecast. In the latter case, we show that the asymptotic bias and mean squared prediction error of the deterministic and Monte Carlo predictors have new elements with respect to the traditional analysis in the absence of data uncertainty.