Statistical Modelling for District-Level Development Planning

The Steering Committee for National Sample Surveys (NSS) constituted by the National Statistical Commission (NSC) recommended that MoSPI should undertake a pilot study on the feasibility for generating the model-based district-level estimates based on the Household Consumption Expenditure Survey (HCES): 2022-23 data. To explore the possibility of generation of district level estimates of Monthly Per Capita Consumption Expenditure (MPCE), a committee was formed under the Chairpersonship of Dr. Mausumi Bose, former Professor, Indian Statistical Institute (ISI), Kolkata with members from the NSS and DES, Uttar Pradesh. The Committee was mandated to estimate MPCE figures for the districts of Uttar Pradesh utilizing model-based techniques, more specifically, Small Area Estimation (SAE) methods. The district-wise MPCE Estimates (in Rs.) of Uttar Pradesh (U.P.) using Model-based Small Area Estimation (SAE) Approach for Rural and Urban Sector are given in Annexure.

Model-based techniques, adopting the Fay–Herriot (FH) and Spatial Fay–Herriot (SFH) models under the Small Area Estimation framework to borrow strength from covariate data and neighboring districts was used to generate reliable estimates of MPCE for the districts of Uttar Pradesh. The detailed estimation methodology along with its benefits and use cases to other identical scenarios have been mentioned in the report available on the website of MoSPI at

https://new.mospi.gov.in/uploads/publications_reports/publications_reports1761641209612_6875c53f-d8eb-4458-be3e-1e12a9e528c9_Compiled_Report_final17092025.pdf

Using the methodology as adopted by the Committee, estimates of MPCE values at sub-state level for the state of Uttar Pradesh, including Baghpat District, have been generated. These estimates of MPCE figures, in turn, would provide an idea of the economic conditions of the households and thereby, facilitating policy decisions for implementation of developmental programmes.

The approach adopted in this study can be extended to other States and other socio-economic indicators. States have been encouraged to adopt similar approaches to fill data gaps and generate district-level estimates, especially where direct survey estimates are weak due to small sample sizes or are not available.

The report has been placed in public domain through MoSPI’s website and is freely accessible by Ministries/Departments and other users.

Annexure

Note: FH denotes Fay–Herriot (FH) model and SFH denotes Spatial FH

This information was given by the Minister of State (Independent Charge) of the Ministry of Statistics and Programme Implementation; Minister of State (Independent Charge) of the Ministry of Planning and Minister of State in the Ministry of Culture Rao Inderjit Singh in Lok Sabha today.