Shapley Regression Based Income Inequality Decomposition of Households in Forest Areas of Southwestern Nigeria

Shapley Regression Based Income Inequality Decomposition

Abstract

The study decomposed income inequality of households in rural southwest Nigeria by Shapley Regression-based decomposition. The Shapley value decomposition approach estimates marginal impact on the level of inequality, which makes it advance over other regression-based inequality decompositions. A proportionate sampling approach was employed, along with a formal survey, to gather information from 450 household heads, of which 430 were deemed valid. Summary statistics tools (specifically percentage, frequency, and mean) and Shapley Regression Base Income Decomposition were used for the analysis. The socioeconomic characteristic showed that maleheaded households constituted 92%, and the majority of the household heads were married. The mean age was 47.63 years, and the average household size was 6.92 persons. Age (0.42), farm size (0.021), distance to market (0.029), and education (0.27) had the highest attributed income share, therefore increasing income inequality. The analytical approach regression result showed that income inequality was exacerbated by being a male-headed household (β=0.6811 high household size (β=0.07) and farm size (β =0.14), while age (β=-0.23) reduced it, the Gini inequality decomposition using Shapley value gave sources that largely explained inequality such as farm size and household size with relative contributions sum up to 25.40%. The weighted marginal contribution results for all the income source variables show that there was a progressive inequality impact, which reduced from level 2 to level 10. This study emphasized that inequality cannot be attributed to a single factor, but rather to a combination of factors. Therefore, stakeholders should not focus solely on a single factor to address the challenges of inequality.