REGRESSION-BASED DECOMPOSITION OF INCOME INEQUALITY IN LAMPUNG PROVINCE INDONESIA

Abstract

This study analyzes the influence of socio-economic and infrastructure factors on gross regional domestic product per capita and measures each factor's contribution to income inequality across 15 districts/cities in Lampung Province during 2020–2025. Using panel data regression with Random Effects Model (REM) and cluster-robust standard errors, the results show that mean years of schooling and internet access have significant positive effects, while poverty rate, open unemployment rate, and electricity access have significant negative effects on GRDP per capita. Applying the Regression-based Decomposition by Fields, all variables collectively contribute 100% to explaining inequality, with open unemployment rate as the largest contributor (16.85%), followed by percentage of poor population (10.40%), electricity access (9.74%), and internet access (9.21%). Notably, mean years of schooling reduces inequality by -4.64%. The overall model explains 41.56% of total GRDP per capita inequality, while the remaining 58.44% reflects residual factors outside the model such as institutional, geographical, and structural conditions. These findings suggest that improving education quality and expanding internet infrastructure can simultaneously boost economic growth and reduce income disparities, while addressing unemployment and poverty remains critical for equitable development in Lampung Province.