Education, household responsibilities, and income in remote border villages

Abstract

This study examines how completed schooling and non-working dependents are associated with reported monthly household income in Nainsid, a remote border village in North Kalimantan, and whether the interpretation changes after household-scale adjustment. A secondary reanalysis of a complete enumeration of household heads applied descriptive statistics, ordinary least squares with heteroskedasticity-consistent standard errors, log-income regression, robust and median estimators, influence diagnostics, and denominator stress tests. Schooling showed a modest positive relationship with income, but the strength and precision of the estimate depended on the estimator and influential observations. Dependents were positively related to total household income, yet this relationship disappeared when income was divided by the number of people explicitly recorded as supported. The study demonstrates that a monetary variable labelled as poverty can reverse substantive interpretation and that unadjusted household income can misrepresent resource adequacy when household composition differs. Rural policy should connect education to employment, productive assets, infrastructure, finance, and markets, while local monitoring should distinguish total income, household composition, and person-level welfare.