Disaggregated Agricultural Output and Economic Growth in Nigeria: Evidence from Distributional Regression Models

Abstract

The role of agriculture in spurring economic growth is often treated as a homogenous industry and modeled under limiting Gaussian conditions. These strategies blur heterogeneity in sub sectors of agriculture and do not reflect the asymmetric and volatile character of growth in resource-dependent economies such as Nigeria. This paper analyzes the impacts of disaggregated agricultural subsector output on the economic growth of Nigeria, and evaluates the ability of both the flexible distributional regression and the traditional normal multiple regression to make better inferences. The study models the real GDP using 60 quarterly data between 2010 and 2024 to determine its relationship between crop, livestock, forestry, and fisheries output. A Generalized Additive Model of Location, Scale and Shape (GAMLSS) is utilized to enable agricultural subsectors to not only influence the conditional mean of the growth, but the variance of the growth and the distributional shape. The performance of the models is compared to a reference Gaussian regression with the information criteria and diagnostic tests. The findings indicate that production of crops is an important driver of growth followed by forestry with a minor contribution. There are weak relationships between livestock and fisheries and aggregate output. The heterogeneous models that use heavy-tailed and asymmetric distributions perform better in comparison to the Gaussian specification, which indicates that the growth process in Nigeria is volatile and asymmetric. The study recommends renewed policy efforts targeted at agricultural modernization, infrastructural development, and sectoral stabilization to enhance agriculture’s contribution to economic growth.