View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Research Papers in Economics AN EMPIRICAL EVALUATION OF SMALL AREA ESTIMATORS Alex Costa1, Albert Satorra2 and Eva Ventura2 1 Statistical Institute of Catalonia (Idescat) Via Laietana, 58 - 08003 Barcelona, Spain E-mail:
[email protected] 2Department of Economics and Business Universitat Pompeu Fabra, 08005 Barcelona, Spain E-mail:
[email protected] ,
[email protected] Abstract This paper compares five small area estimators. We use Monte Carlo simulation in the context of both artificial and real populations. In addition to the direct and indirect estimators, we consider the optimal composite estimator with population weights, and two composite estimators with estimated weights: one that assumes homogeneity of within area variance and squared bias and one that uses area-specific estimates of variance and squared bias. In the study with real population, we found that among the feasible estimators, the best choice is the one that uses area-specific estimates of variance and squared bias. Key words: Regional statistics, small areas, root mean square error, direct, indirect and composite estimators. AMS classification (MSC 2000): 62J07, 62J10, 62H12. Acknowledgements: The authors are grateful to Xavier López, Maribel García, Cristina Rovira and Antoni Contel for their help at several stages of this paper. We are also indebted to Nick T. Longford for detail comments on a previous version of this paper. 1 Introduction Official statistics is faced with the need to generate estimates for small administrative units; while working with relatively small samples and within stringent budgetary limit.