District-level Estimates of Unemployment Rates in Odisha

This article describes the possibilities of using an alternative method, such as small area estimation, for generating district-level unemployment estimates with higher precision. The SAE method is applied to generate the unemployment rate of different districts of Odisha combining the Periodic Labour Force Survey 2018–19 data of the National Sample Survey Office and the auxiliary variables from other secondary data sources.

An immediate fallout of the COVID-19 pandemic was the shrinkage in the economy and the rise in unemployment or underemployment. According to the Centre for Monitoring Indian Economy (CMIE), the unemployment rate (UR) slowed from 6.7% during October 2020 from the peak rate of 23.5% in April 2020 after the nationwide lockdown. The latest Periodic Labour Force Survey (PLFS) 2018–19 of the National Sample Survey Office (NSSO) reported that the usual status UR, including principal status (ps) and subsidiary status (ss), in Odisha was 7%. The urban unemployment was 12.7%, and rural unemployment 6%.

Planners and policymakers need statistical data for efficient policymaking at the district level. Despite its importance, the URs at the lower administrative levels like the district are generally unavailable (Anjoy et al 2019). The sample size of large-scale surveys like PLFS, which provides unemployment data at the national/state/regional levels, is small and cannot provide reliable district-level estimates with adequate precision. Some alternative methods such as small area estimation (SAE), which borrows strength (information) from other related areas to arrive at reliable estimates with lower standard errors at the small area level, are becoming important in survey sampling and is emerging as a viable and cost-effective solution (Chandra et al 2011).