Understanding Brazilian Unemployment Structure: A Mixed Autoregressive Approach

The aims of this paper are estimate and forecast the Non-Accelerating Inflation Rate of Unemployment, or NAIRU, for Brazilian unemployment time series data. In doing so, we introduce a methodology for estimating mixed additive seasonal autoregressive (MASAR) models, by the Generalized Method of Moments (GMM). Furthermore, in order to cover a lack in econometric literature, an asymptotic theory for the Yule-Walker estimators of autoregressive parameters is developed. The paper provides some insights on estimating MASAR models when one of its component has a possible unit root. The obtained results are consistent to the literature and produce reasonable forecasts for NAIRU.