Assessment of Socioeconomic Determinants of Voting Patterns: A Spatio-Temporal and Multinomial Logits Analysis of the 2023 Nigerian Presidential Election

Abstract: In Nigeria, electoral behavior is often interpreted through ethno-religious views and regional allegiances, without empirically assessing the influence of socioeconomic indicators such as health, income, education, and other deprivations on voter behavior. This study investigates the voting pattern in the 2023 Nigerian presidential election and previous cycles using spatio-temporal and multivariate analysis. It examines whether support for some candidates was more substantial in states with higher Human Development Index (HDI) and whether this alignment impacts governance quality and macroeconomic performance post-election. Socioeconomic data were obtained from the Global Data Lab and the Nigerian Bureau of Statistics (NBS) for the year preceding each election, while presidential vote results were sourced from the Independent National Electoral Commission (INEC). The results show that the Labour Party (LP) dominated states with high socioeconomic indices, accounting for about 30-53% of the variance in voting patterns of LP and People's Democratic Party (PDP), while the All Progressive Congress (APC) variance is less explained by these factors. Multinomial logits based on HDI is used to predict party win probabilities; the model predicted about 60% wins accurately. Comparative analysis of four presidential cycles revealed that in 2011, the winner had an HDI-vote correlation of 0.44, with improved macroeconomic indices post-election. In contrast, 2015, 2019, and 2023 saw negative correlations of -0.38, -0.43, and -0.34, respectively, alongside macroeconomic decline. The findings suggest that socioeconomic development shapes political preferences, promotes issue-based politics, and supports quality leadership; therefore, strengthening education, healthcare, and poverty reduction should be prioritized to enhance citizens' well-being and build an informed electorate.