Abstract:
|
The paper proposes an adaptive expectations model to jointly project the main macroeconomic aggregates of Argentina. The model is essentially an autoregressive vector with exogenous variables (VARX) with the novelty that its parameters are not estimated by OLS but by linear programming in order to incorporate equality and inequality constraints and prior information as well. These constraints are intended to bound the estimated parameters within the parametric space suggested by the economic theory, while previous estimates are introduced to minimize the discrepancies between projections made with homologous series from different base years. The study concludes that projections obtained from the adaptive expectations model fit well to macro aggregates of Argentina's National Accounts System, although they do not always outperform those made with specific ARIMA models for each aggregate.
|