Small Area Estimation of Poverty Incidence with Sampling Error Variances Through Generalized Variance Function

Norberto Espejo Milla

doi:10.11648/j.ajtas.20170602.11

Abstract

Small area estimation based on area level models, particularly the EBLUP method, typically assumes that sampling error variances of the direct survey small area estimates are known. In practice, the sampling error variances are unknown. This paper generates EBLUP estimates of poverty incidence when the sampling error variances are estimated using the generalized variance function (GVF) approach. The precision of the EBLUP estimates is determined using a modified version of the Prasad-Rao MSPE estimator. The modification is made by adding an extra term that would account the uncertainty associated with estimating the sampling error variances. The performance of the modified Prasad-Rao estimator relative to the commonly used Prasad-Rao estimator is evaluated through a simulation study. Results have shown that the modified Prasad-Rao MSPE estimator has relatively greater bias than the commonly used Prasad-Rao MSPE estimator, particularly for small samples. A slight gain in precision is observed when using the modified PR MSPE estimator, especially for large samples. Moreover, the findings imply that estimating sampling error variances using GVF models can be a very useful strategy in the application of EBLUP small area estimation, most particularly in poverty incidence estimation.

Highlights

In the Philippines, nationwide surveys such as the Family Income and Expenditure Survey (FIES), Labor Force Survey (LFS), and the National Nutrition Survey (NNS) are the principal sources of official statistics of the country
Modeling sampling error variances using generalized variance function (GVF) approach has been shown to be a potential technique for use in empirical best linear unbiased prediction (EBLUP) estimation of poverty incidence
The technique has produced EBLUP estimates which are statistically lower than the EBLUP estimates obtained assuming known sampling error variances

Summary

Introduction

In the Philippines, nationwide surveys such as the Family Income and Expenditure Survey (FIES), Labor Force Survey (LFS), and the National Nutrition Survey (NNS) are the principal sources of official statistics of the country These surveys contain regions as sampling domains since it is timeconsuming and expensive to conduct the surveys with domains at the provincial or city levels. Aside from purely computing these statistics at levels smaller than the domain of the survey, say, at the provincial level, it is important to ascertain that these statistics are precise so that proper targeting of the right beneficiaries of these programs is achieved In this way, government laws and programs can efficiently be delivered to those localities or even households that are really in greatest need. The DSWD used the 2003 Small Area Estimates of poverty incidence generated by the National Statistical Coordination Board (NSCB) as one of the criteria for identifying eligible households for its Conditional Cash Transfer Program, otherwise known as Pantawid Pamilyang Pilipino Program (4Ps)

Methods

Results

Conclusion