Bias nonmonotonicity in stochastic difference equations

Paper number: 95/12

Paper date: October 18, 1995

Year: 1995

Paper Category: Discussion Paper


Karim M Abadir


Kaddour Hadri*


We show that the bias of estimated parameters in autoregressive models can increase as the sample size grows. This unusual result is due to the effect of the initial sample observations that are typically neglected in theoretical asymptotoc analysis, in spite of their empirical relevance. Implications for practical economic modelling are considered.

JEL Classification numbers: C22.
Keywords: Economic Time Series, Near-Nonstationarity, Estimator Biases.

Corresponding author: Karim Abadir, Department of Economics, Amory Building, University of Exeter, Exeter EX4 4RJ, England.

*We would like to thank Jane Black, David de Meza, Ben Lockwood Gareth Myles and Garry Phillips for useful comments on this work.