Prediction in the oneway error component model with serial correlation
Academic Article
Overview
Identity
Additional Document Info
Other
View All
Overview
abstract
This paper derives the best linear unbiased predictor for a oneway error component model with serial correlation. A transformation derived by Baltagi and Li (1991) is used to show how the forecast can be easily computed from the GLS estimates and residuals. This result is useful for panel data applications which utilize the error component specification and exhibit serial correlation in the remainder disturbance term. Analytical expressions for this predictor are given when the remainder disturbances follow (1) an AR(1) process, (2) an AR(2) process, (3) a special AR(4) process for quarterly data, and (4) an MA(1) process. Copyright 1992 John Wiley & Sons, Ltd.