Public Expenditure On Education And Health And Its Impact On Inequality: Household Evidence From Lucknow District, Uttar Pradesh

Abstract

Public spending on education and health is the largest instrument of redistribution available to an Indian district 
administration, yet its distributional consequences are almost never measured at the district scale at which the 
services are actually delivered. This paper estimates how in-kind provision of schooling and medical care alters 
consumption inequality in Lucknow district, Uttar Pradesh. A stratified two-stage survey of 1,456 households 
drawn from 108 primary sampling units across the eight rural development blocks and four municipal zones is 
used to value the education and health services each household actually consumes, at government unit cost, and 
to add that value to monthly per capita consumption expenditure. In-kind provision is worth Rs 363 per capita 
per month, or 6.02 per cent of mean consumption, and it is strongly pro-poor in absolute terms: the poorest 
quintile receives Rs 500, equivalent to 19.54 per cent of its own consumption, against Rs 196 and 1.70 per cent 
for the richest. The concentration indices are negative for both services, and the Kakwani indices of −0.487 for 
education and −0.461 for health confirm that each is distributed more equally than consumption itself. Valuing 
the services reduces the Gini coefficient from 0.2965 to 0.2711, a fall of 2.54 points or 8.55 per cent, of which 
reranking absorbs only a negligible part. Recentred influence function regressions with stratum fixed effects and 
standard errors clustered by sampling unit confirm the association, while an instrumental-variables specification 
using distance to the nearest facility is too imprecise to be informative. A Monte Carlo replication of 200 
independent survey draws establishes which of these findings are robust and which are artefacts of a single 
sample.