The National Statistics Office (NSO), under the Ministry of Statistics and Programme Implementation (MoSPI), has released a new study on model-based district-level estimates derived from the Household Consumption Expenditure Survey (HCES) 2022–23 for Uttar Pradesh. The report, now available on the MoSPI website, marks a step towards more localised, data-driven policymaking.

The NSO conducts large-scale household surveys across diverse socio-economic themes to generate reliable statistical inputs for policymaking. Among these, the HCES plays a vital role by capturing data on household consumption patterns, living standards, and demographic characteristics at national and state levels.

Aim of the Study

The study was undertaken following the National Statistical Commission’s Steering Committee recommendation to pilot model-based estimation techniques for generating district-level insights. A dedicated committee chaired by Dr Mausumi Bose, Former Professor at the Indian Statistical Institute (ISI), Kolkata, was constituted to explore the feasibility of estimating Monthly Per Capita Consumption Expenditure (MPCE) for each district in Uttar Pradesh using HCES data.

The project received technical support from the NSO and the Directorate of Economics and Statistics (DES), Government of Uttar Pradesh.

While HCES data provides robust estimates at national and state levels, it often lacks district-level precision due to limited survey samples. To address this gap, the study adopted a model-based approach, testing whether statistical modelling could generate accurate, cost-effective estimates for smaller administrative units.

Methodology

The research employed a statistical technique called Small Area Estimation (SAE), which enhances data accuracy for smaller regions by combining survey data with auxiliary administrative information. This approach “borrows strength” from related datasets, improving the stability of estimates where direct sampling is insufficient.

The study utilised two types of statistical models — Fay–Herriot (FH) and Spatial Fay–Herriot (SFH) — and incorporated auxiliary data such as:

Number of old-age pension beneficiaries

Number of Ayushman Bharat (PM-JAY) patients

Number of domestic LPG connections

Number of Antyodaya food scheme beneficiaries

Key Findings

The top five rural districts with the highest average MPCE were:

Bagpat

Saharanpur

Gautam Buddha Nagar

Meerut

Ghaziabad

In urban areas, the leading districts were:

Gautam Buddha Nagar

Gonda

Ghaziabad

Bagpat

Lucknow

The study showed that model-based estimation can be a cost-effective and scalable solution for generating district-level statistics using state-level survey data.

Conclusion

The findings reaffirm the potential of statistical modelling as a reliable tool for filling data gaps and improving local-level governance. By providing district-specific insights, the method enables policymakers to design targeted welfare programmes, monitor living standards, and reduce regional inequalities.

The success of this pilot in Uttar Pradesh sets a precedent for extending the model-based approach to other states and socio-economic indicators, such as employment, health, and poverty, advancing India’s commitment to data-driven policymaking and sustainable development.