FORECASTING IN LEAST ABSOLUTE VALUE REGRESSION WITH AUTOCORRELATED ERRORS - A SMALL-SAMPLE STUDY

Least absolute value (LAV) regression is a robust alternative to ordin ary least squares (OLS) and is particularly useful when model disturba nces follow distributions that are nonnormal and subject to outliers. The performance of the OLS estimator when the disturbances are autocor related has been studied extensively, but the performance of the LAV e stimator in the presence of serial correlation is less well establishe d. In this research, we study the forecasting performances of OLS- and LAV-based models for simple time series regression when the errors ar e autocorrelated. Monte Carlo simulation methods are used to compare t he forecasting accuracies of the different models. A least absolute va lue analogue of the Prais-Winsten correction possesses an appealing ro bustness for the context under consideration.