Bias reduction in autoregressive models

It is well known that standard estimators of an AR(p) model are biased in f inite samples, yet little is done in practice to remove the bias. Apparentl y small biases have important implications for the estimation of the impuls e response function, which is a nonlinear function of the original coeffici ents. This note shows how to obtain estimators adjusted for first order bia s based on Stine and Shaman's fixed point characterisation of the bias (Sti ne, R.A., Shaman, P., 1989. A fixed point characterisation for bias of auto regressive estimators. Annals of Statistics 17, 1275-1284). (C) 2000 Elsevi er Science S.A. All rights reserved. JEL classification: C22.