The Nonlinear Impact of Education, Capital, and Labor on Regional Income Distribution in Indonesia

This study explores how education, physical capital, and labor interact in shaping income inequality across Indonesia’s 34 provinces between 2004 and 2023. Traditional linear models often miss the nuances here—they tend to assume that these factors affect all regions in the same way, regardless of their stage of development. To dig deeper, this research applies a Generalized Additive Model (GAM), which captures nonlinear patterns and thresholds that change depending on regional conditions. The results paint an interesting picture: primary, junior, and senior high school education consistently help narrow inequality, while tertiary education follows an inverted U-shaped trend. In other words, higher education can initially widen income gaps before eventually helping to reduce them as more graduates enter the workforce. Physical capital shows diminishing equalizing power—it’s most effective in capital-scarce regions and less so as investment levels grow. Labor, on the other hand, forms a U-shaped relationship with inequality: it reduces disparities up to a point, after which its benefits start to taper off once regions hit their absorptive limits. Overall, the GAM model performs notably better than standard linear fixed-effect approaches, explaining 82.4% of the deviance compared to 68.1%. These findings highlight the need for more tailored policies that reflect each region’s stage of economic maturity—focusing on education type, capital accumulation, and labor development in a way that fits local realities.