© The Indian Express Pvt Ltd
Tags:
Journalism of Courage
Sonbhadra, the second-largest district in Uttar Pradesh, has some of the lowest spenders in the state, with rural consumption in the district just Rs 2,337 per month in 2022-23 – almost half that of rural Gautam Buddha Nagar. Meanwhile, in urban areas of Gautam Buddha Nagar district, which houses the city of Noida and is part of the Delhi National Capital Region, the monthly per capita consumption expenditure is almost Rs 10,000 — by far the highest of any district in Uttar Pradesh. In contrast, residents of urban Ballia, Uttar Pradesh’s easternmost district, spent just Rs 2,575 per month.
The findings are the result of a pilot study by the Ministry of Statistics and Programme Implementation (MoSPI), which used data from the Household Consumption Expenditure Survey (HCES) for 2022–23 to arrive at district-level consumption expenditure estimates for Uttar Pradesh amid “growing demand” for such data “to support local planning and monitoring of welfare schemes”.
As per MoSPI’s HCES, the monthly per capita consumption expenditure of rural Uttar Pradesh in 2022-23 was Rs 3,191. For urban areas, it was Rs 5,040. The corresponding figures for India as a whole were Rs 3,773 for rural and Rs 6,459 for urban areas.
The release of the pilot study comes weeks after meetings between officials from MoSPI and the UP government. On October 14, Chief Minister Yogi Adityanath and other state officials met MoSPI Secretary Saurabh Garg and Rajiva Laxman Karandikar, Chairman of the National Statistical Commission, among others. The meeting focussed on strengthening Uttar Pradesh’s statistical system to aid it in its objective of becoming a $1 trillion economy.
At the meeting of the National Statistical Commission with Uttar Pradesh’s Directorate of Economics and Statistics, state officials reiterated their “commitment to evidence-based planning and data-driven governance”, as per a government statement, with the framework of the One Trillion Dollar Economy Mission also being presented at the interaction.
District-level numbers
A key difference between the district-level numbers and state and national estimates of monthly per capita expenditure as per the HCES is that while the latter is found directly from a survey conducted by speaking to households, the former is derived from a model. This is because while the number of households contacted during the HCES at the state and all-India level is sufficiently large to result in reliable estimates, the sample size is relatively small when it comes to individual districts.
For instance, more than 2.6 lakh households were surveyed by MoSPI for the 2022-23 HCES. Of these, only around 30,000 were from Uttar Pradesh. The numbers get even smaller for any one district.
As such, a committee led by Mausumi Bose, formerly a professor at Indian Statistical Institute, Kolkata, used a statistical method called ‘Small Area Estimation’ that combines survey data with others such as the number of pensioners, patients under Ayushman Bharat scheme, and domestic LPG connections, among others, to calculate the monthly per capita consumption expenditure for each district of Uttar Pradesh.
“This study highlights how statistical models can help fill data gaps where direct survey results are limited, providing policymakers and planners with valuable insights to design and evaluate welfare programmes, monitor living standards, and reduce regional inequalities. The approach adopted in this study can be extended to other states and other socio-economic indicators such as employment or poverty, thereby advancing data-driven governance and local-level planning,” a statement by the statistics ministry said on Monday.
Choosing between estimates
With the ‘direct’ estimates of district-level monthly per capita consumption expenditure generated by the HCES deemed unreliable because of the small number of households surveyed in each district, two statistical models were used to find out how much people were spending each month: Fay-Herriot (FH) and Spatial Fay-Herriot (SFH) models.
According to the study, both the models resulted in more reliable estimates, with the SFH model leading to even better results especially for rural areas. The figures for district-level monthly per capita consumption expenditure mentioned here are from the SFH model.
