MULTIVARIATE ANALYSIS OF SECTORAL INFLUENCES ON HOUSEHOLD FOOD CONSUMPTION EXPENDITURE USING PRINCIPAL COMPONENT AND REGRESSION MODELING

Abstract

In this study, the principal component analysis (PCA) and multiple regression were adopted to examine the sectoral determinants of household food consumption expenditure (as proxy to food security) in Nigeria. Annual data from the Central Bank of Nigeria Statistical Bulletin spanning 1996 to 2021 were used. The results showed that the correlation analysis revealed strong positive relationships among the sectoral variables, implying the presence of multi-collinearity. The outcomes of PCA indicate that the first principal component (PC1) represents 93.58% of the total variance, whereas the second and third components are 3.96 and 1.56, and this fact confirms the fact that the data is characterized by only one underlying economic structure. The PCA loadings show that all sectors have a positive contribution to PC1 with average contribution of approximately 13%, which implies that sectors have a balanced and interconnected influence in household’s consumption. Nevertheless, the energy sector pushes the second major element by nearly 60%, which shows a structural role. The individuals plot also indicates three different clusters, showing the difference in time on sectoral performance. The regression findings demonstrate that the relationship between agriculture and household consumption expenditure is positive and statistically significant ( = 0.5886, p < 0.05) and water supply and ICT are not significant. The model describes 94.98% of the change in household consumption expenditure ( = 0.9498), which shows a great fit in the model. This result shows the sectoral interactions, especially agriculture, strongly influence household consumption, supporting policies that enhance welfare and food security.