Abstract: Paper No. 15
When Random Group Effects are Cross-Correlated: An Application to Elderly Migration Flow Models
Karen Smith Conway and Andrew J. Houtenville
Abstract: Incorporating random group effects has proven important to making correct statistical inferences about factors that only vary across groups. We note that it is possible to have more than one random effect in models using cross-sectional data and that these random effects could be correlated, unlike in the typical panel data situation. Extending the standard multiple random effects model in this way is greatly simplified by using the two-step estimator we develop. Our application to an elderly migration flow model provides an intuitive example of cross-correlated random group effects and demonstrates the ease of our estimator, as well as highlighting the empirical importance of controlling for random effects.
For more information on ordering a hard copy of this paper, please contact the Publications Officer, Center for Policy Research, 426 Eggers Hall, Syracuse University, Syracuse, New York 13244-1020 or e-mail our Publications Officer at email@example.com.