Small area estimation under unit-level generalized additive models for location, scale and shape

TL;DR

This paper introduces SAE-GAMLSS, a flexible small area estimation model that relaxes distributional assumptions and improves estimation accuracy of household economic indicators across regions.

Contribution

The paper develops SAE-GAMLSS, a novel small area estimation method based on GAMLSS, allowing covariate-dependent distributional parameters and providing a bootstrap MSE estimation approach.

Findings

SAE-GAMLSS outperforms EBLUP in simulated scenarios.

It reveals regional differences in household consumption in Italy.

The North-South divide does not apply to foreigners.

Abstract

We propose a Small Area Estimation model based on Generalized Additive Models for Location, Scale and Shape (SAE-GAMLSS), for the estimation of household economic indicators. SAE-GAMLSS release the exponential family distributional assumption and allow each distributional parameter to depend on covariates. A bootstrap approach to estimate MSE is proposed. The SAE-GAMLSS estimator shows a largely better performance than the well-known EBLUP, under various simulated scenarios. Based on SAE-GAMLSS per-capita consumption of Italian and foreign households in Italian regions, in urban and rural areas, is estimated. Results show that the well-known Italian North-South divide does not hold for foreigners.