Village level poverty rates are needed as a consideration for allocating village funds. The national socio economic survey
samples are designed to estimate poverty rates in province and distric level. Direct estimate for calculating estimates of village level
poverty rates does not have a good precision due to small sample sizes. Small Area Estimation (SAE) technique is used to produce a
good precision with small sample sizes. The estimates of poverty rates should also be produced for non sampled area and when no poor
are included in the sample. We propose zero inflated beta model because poverty rates takes value in the intervals [0,1). Clustering
technique is used to acommodate random effect area for non sampled area. The purpose of this research is to estimate poverty rates on
village level in Langsa Municipality. The result showed that estimates poverty rates on village level with zero inflated beta model is
better than direct estimates.

Published In:IJCSN Journal Volume 6, Issue 6

Date of Publication : December 2017

Pages : 812-819

Figures :01

Tables : 06

Meita Jumiafrtanti : currently pursuing masters degree program in
Applied Statistics in Bogor Agricultural University

Indahwati : is lecturer at Department of Statistics , Bogor
Agricultural University, Indonesia. Her main interest is in Mixed
Model and Small Area Estimation.

Anang Kurnia : is lecturer at Department of Statistics , Bogor
Agricul-tural University, Indonesia. His main interest is in Small
Area Estimation.

Clustering, Poverty Rates, Small Area Estimation, Zero Inflated Beta Model

The conclusion is zero inflated beta model in the small
area estimation produce a better estimator because no
value is 0 at village level, whereas in the direct estimator
there is a value of 0. Zero inflated beta model in the small
area estimation to estimate non sampled area also produce
good estimator values while still considering the fixed
influence and random diversity of village areas. However,
the result of zero inflated beta model in the small area
estimation with the weighted average in village level still
has a value under the direct estimator of the poverty rates
of Langsa Municipality level, so it is necessary to calibrate
the model, this is probably due to the value in the response
variable contains a lot of value 0.

[1] Anisa. R, “Kajian Pengaruh Penambahan Informasi
Gerombol Terhadap Hasil Prediksi Area Nircontoh
(Studi Kasus Pengeluaran per Kapita Kecamatan di Kota
dan Kabupaten Bogor)”, M.Si. thesis, Department of
Statitics, Bogor Agricultural University, Bogor,
Indonesia, 2014.
[2] [BPS] Badan Pusat Statistik. Penghitungan dan Analisis
Kemiskinan Makro Indonesia Tahun 2016. Jakarta:
Katalog BPS, 2016
[3] Ferrari. SLP, and Cribari-Neto. F, “Beta Regression For
Modelling Rates and Proportions”, Journal of Applied
Statistics, Vol. XXXI, No. VII, 2004, pp. 799-815.
[4] Foster. J, Greer. J, and Thorbecke. E, “A class of
decomposable poverty measures”, Econometrica, 52,
1984, pp.761-766.
[5] Kurnia. A, and Notodiputro. KA, “Penerapan Metode
Jackknife dalam Pendugaan Area Kecil”, Forum
Statistika dan Komputasi, Vol. XI, No. I, 2006, pp. 12-
15.
[6] Ospina. R, and Ferrari. SLP, “Inflated beta
distributions”, Statisticals Papers, 51, 2010, pp.
111–126.
[7] Rao. JNK, Small Area Estimation. New York (US):
John Wiley & Sons. 2015
[8] Swearingen CJ, Castro MSM, Bursac Z. “Modeling
percentage outcomes: the %beta_regression macro”,
SAS Global Forum 2011: Statistics and Data Analysis
Paper, 335, 2011, pp.1-12.