The Paradox of Capital Expenditure Effectiveness on GRDP Per Capita: Poverty, HDI, and Fiscal-Social Decoupling in Jambi Province

Abstract

This study aims to examine the effect of capital expenditure, poverty, and the Human
Development Index (HDI) on GRDP per capita across 11 regencies/municipalities in Jambi
Province during 2017–2024. Secondary panel data (88 observations) from Statistics Indonesia
(BPS) Jambi Province and the Directorate General of Fiscal Balance were analyzed using panel
data regression, with model selection through the Chow and Hausman tests. The tests establish
the Fixed Effect Model (FEM) as the best model (Chow: Prob. Cross-section Chi-square =
0.0000; Hausman: Prob. Cross-section random = 0.0474). Simultaneously, capital expenditure,
poverty, and HDI significantly affect GRDP per capita (Prob. F-statistic = 0.000000), with high
explanatory power (R-squared = 0.703572; Adjusted R-squared = 0.615086). Partially,
however, none of the three variables is statistically significant (Prob. t of 0.9353, 0.7573, and
0.6815, respectively), with capital expenditure and poverty showing negative coefficients and
HDI a positive one. This pattern suggests that the model's high explanatory power is driven
largely by cross-sectional fixed effect time-invariant structural characteristics of each
regency/municipality rather than by the annual dynamics of the three policy variables
themselves. These findings are interpreted through the lens of Fiscal-social decoupling: a
disconnection between annual fiscal-social instruments and regional economic outcomes, in
light of Jambi's economic structure, which remains dominated by primary and commodity-based
sectors. The novelty of this study lies in operationalizing decoupling through the pattern of
partial insignificance amid a strong joint model, read as policy evidence of the dominance of
regional structural factors rather than mere model failure. The article recommends shifting
capital expenditure from an input-based to an outcome-based approach, strengthening
productive infrastructure linked to leading sectors, integrating poverty-reduction and HDI
programs with regional economic strategy, and using economic-benefit indicators in local
budget evaluation. These findings should be read within the limitations of a model that has not
yet incorporated structural variables, private investment, commodity prices, or lagged capital
expenditure.