Skip to main content Accessibility help
  • Cited by 7
  • Print publication year: 1999
  • Online publication date: September 2009

10 - Prediction from the regression model with one-way error components


Econometrics had its origin in the recognition of empirical regularities and the systematic attempt to generalize these regularities into “laws” of economics. In a broad sense, the use of such “laws” is to make predictions – about what might have been or what will come to pass. Econometrics should give a base for economic prediction beyond experience if it is to be useful. In this broad sense it may be called the science of economic prediction.

Lawrence R. Klein (1971)


Following the work of Balestra and Nerlove (1966) the regression model with error components, or variance components, has become a popular method for dealing with panel data. A summary of the main features of the model, together with a discussion of some applications, is available in Hsiao (1986), Mátyás and Sevestre (1992), Maddala (1993), and Baltagi (1995). However, relatively little is known about prediction from the model. Assuming that all the regression parameters and the error process parameters are known, the form of the optimal (in the sense of minimum MSE predictor) has been obtained by Wansbeek and Kapteyn (1978) and Taub (1979). This was extended to the case of serially correlated error components by Baltagi and Li (1992).

This chapter investigates some potentially important practical problems associated with prediction from the error components regression model.