Assessment of non-linearity in calorie–income relationship in Pakistan

Abstract

Read online

This article considers the issue of assessing non-linearity in the relationship between calorie consumption and income using non-parametric and semi-parametric approaches. These methodologies are implemented on the cross-sectional household survey data conducted in Pakistan in 2010–2011. This framework takes account of the heterogeneity among families and potential non-linearity in the relationship. The findings show that the calorie–income elasticity is considerable and statistically significant across estimating methodologies. The results also demonstrate that the elasticity is larger for the substantially poorer households of the sample. By incorporating the explanatory variables in a manageable way in the parametric section of regression procedures, the semi-parametric analysis also reveals a slight increase in calorie response to increases in income at various income levels.

Keywords

• calorie–income
• non-parametric regression
• semi-parametric regression
• single-index model
• non-linear elasticities
• Pakistan