INTRODUCTION
1.1. BACKGROUND
Bangladesh Bureau of Statistics (BBS) conducted the first round of Household Expenditure Survey (HES) in 1973. The latest i.e. the 17th round of HIES was held in 2022. National and Divisional level (rural and urban) poverty Head Count Rates (HCR) are generated directly from the HES/HIES survey datasets. However, the District and Upazila level poverty rates are highly demanded by the policy makers, development partners and the researcher’s community too. To meet the stakeholder’s high expectations, BBS started publishing the District and Upazila poverty pictures by using the Small Area Estimation (SAE) technique with the collaboration of WFP and WB since 2000. However, the survey figures show that the poverty has undergone a profound shift from a high 48.9 percent in 2000, the poverty rate plummeted to 18.7 percent by 2022.1 Despite these strides, marked disparities persist across different geographical areas and communities. Understanding these spatial disparities is crucial for formulating effective policies tailored to address these multifaceted challenges. The ‘Poverty Map of Bangladesh 2022’ provides a detailed poverty distribution across the country, embodying Bangladesh’s enduring commitment to poverty alleviation. It is worth to mention here that the only exception was HIES 2016 where the National, Divisional and also the District HCRs were given directly from the survey and the Upazila level figures were produced through SAE method.
The traditional household surveys are invaluable for assessing poverty at national or large regional levels.2 Yet, their capacity to capture the nuanced disparities in smaller or more specific areas often falls short due to many reasons including limitations in sample size. In areas where only a few households are surveyed, the results may not accurately reflect the broader local conditions, leading to a potentially skewed understanding of poverty and its distribution. The SAE techniques are specifically developed to address these shortcomings by enhancing the precision of poverty estimates for smaller geographic areas or specific demographic subgroups, that traditional surveys cannot capture due to smaller sample size.
The SAE achieves this enhanced accuracy by integrating detailed survey data with auxiliary information including census data, administrative records, and potentially satellite imagery or mobile data. This methodology allows for ‘borrowing strength’ from related areas or groups, significantly increasing the reliability of the estimates where direct survey data is sparse. For instance, SAE leverages demographic and economic patterns identified in the census—which includes every household in the country—to refine and adjust poverty estimates derived from survey data.
In the development of the ‘Poverty Map of Bangladesh 2022’, SAE techniques were utilized, capitalizing on data from the Household Income and Expenditure Survey (HIES) 2022 and the Population and Housing Census (PHC) 2022. This approach facilitates the estimation of poverty levels down to the district and upazila levels, offering a granularity that surpasses the divisionlevel estimates typically provided by HIES 2022. The Bangladesh Bureau of Statistics (BBS), in collaboration with two international partners i.e. the World Bank (WB) and the World Food Programme (WFP), played a vital role in spearheading the production of the 2022 poverty maps.
Such detailed mapping of poverty at lower sub-national administrative units is crucial for both government and non-government organizations to allocate resources and taking interventions more effectively. By pinpointing areas of acute need and monitoring progress over time, these maps serve as a foundational tool for targeted poverty alleviation strategies. This ensures that efforts are concentrated where they are most important, promoting equitable development across diverse communities. Furthermore, these detailed measures provide policymakers with a robust mechanism to assess the effectiveness of their policies, particularly in tracking and monitoring the Sustainable Development Goals (SDGs) to be achieved by 2030.