Food Price and Households' welfare comparisons: Evidence from Household Panel Data in Rural Pakistan

Abstract

This study analyzes static and dynamic welfare impacts of real food price changes across various income and land holding groups, low price, and high price periods as well as across region. This study uses panel data set, Pakistan Rural Household Survey, PIDE, with the objective to analyze dynamic consumption behavior. All three rounds (2001, 2004 and 2010) are used. This research work makes three main contributions in literature. First, this study uses QUAIDS demand model to estimate demand functions for eight food groups'cereals, meat, milk and milk products, vegetables, cooking oil, pulses and other food. The system of equations is estimated by NLSUR using the result of Lagrange Multiplier, and of test on the coefficient of squared expenditure term, which supports the inclusion of squared expenditure term in food demand model. This study tests for endogeneity and uses control function for its correction. The problem of observed zero expenditure is dealt with a two-steps procedure. Second, it decomposes overall sample into low price and high price periods, and uses expenditure and price elasticity coefficients from QUAIDS to analyze static and dynamic welfare effects of real food price changes. The compensated variation (CV), money metric loss from food price changes, is estimated for all subsamples: upper and bottom income quintiles, landless, small and large landowners and Sindh and Punjab. Third, this study compares the findings of this study on welfare implications of food price changes based on QUAIDS with the earlier research, which used other specifications of the demand system. In addition, this study analyzes the food insecurity of rural household by using various indicators across periods, and for the entire sample. The indicators used in the study are per adult equivalent calorie, vii calorie availability per capita, calorie from staple food, and percentage share of food expenditures. The findings indicated that dynamic welfare impacts of real food price changes are lower than static welfare impacts in all subsamples. The static and dynamic welfare impacts are higher for poorest in the income group and for landless in the land holding group. Across regions, static and welfare impacts of real food price changes are smaller in Punjab as compared to Sindh. Across periods, money metric losses from food price changes are significantly larger in high price period than low price period. Households are food insecure on the medium scale in the sample period. All indicators, except percentage share of food, worsen in high price period implying negative welfare impact of food price changes on food security. The findings of this study, based on appropriate methodology, answers many questions raised by policy makers on food insecurity.