Penerapan Geographically Weighted Panel Regression dan Data Envelopment Analysis dalam Pemodelan Kemiskinan di Kalimantan Timur

Abstract

Indonesia currently still needs to focus on achieving sustainable development goals agreed by all countries in the world. Indonesia presently ranks 82nd out of 163 nations in terms of SDG accomplishment, indicating that there is still plenty of potential for improvement. One of the goals that hasn't been accomplished is ‘no poverty’. Regarding the poverty cases, among all province in Indonesia, East Kalimantan is important to be analyzed, because Penajam Paser Utara and Kutai Kartanegara in East Kalimantan are scheduled to become Indonesia's next capital, Nusantara. The goal of this research is to investigate the variables that influence poverty in East Kalimantan and determine the effectiveness of poverty alleviation in the regencies/cities in East Kalimantan. This research used indicator data of poverty from 2019-2021 retrieved from Statistics Indonesia. This research use spatial panel data analysis regression method or Geographically Weighted Panel Regression (GWPR) and Data Envelopment Analysis (DEA). In GWPR model, this research compared adaptive gaussian, adaptive bisquare, adaptive exponential, fixed gaussian, fixed bisquare, and fixed exponential kernel. The findings of this investigation revealed that fixed exponential is the kernel that has lowest AIC and the highest adj-𝑅2. The variables that determine poverty of regencies/cities in East Kalimantan are expenditure per capita, life expectancy, and number of village with higher education facilities. Furthermore, according to DEA, only three cities were effective in addressing poverty: Mahakam Ulu, Paser, and Penajam Paser Utara.