Despite having long term efforts, poverty is an important and persistent social issue in India. Existing data based on socio-economic surveys produce state and nationally representative poverty estimates but cannot be used directly to generate reliable disaggregate or local level estimates. The state and national level estimates often mask the variations at the local level which in turn restricts the effective implementation of policies related to poverty alleviation locally within and between administrative units. This paper uses the Household Consumer Expenditure Survey data of NSSO and link with the Population Census data to produce the reliable district-level estimates of poverty incidence in the rural areas of West Bengal in India. In particular, small area estimation (SAE) method is explored to generate reliable district-level poverty estimates. The results clearly indicate that the district-level estimates generated by model-based SAE method are precise and representative. A map showing how poverty incidence varies by district across the State of West Bengal is also produced. The estimates generated from this research are useful for meeting the data requirements for policy research and strategic planning by different international organizations and by Departments and Ministries in the Government of India.

尽管进行了长期的努力，贫困仍然是印度一个重要而持久的社会问题。基于社会经济调查的现有数据可得出州和全国代表性的贫困估计，但不能直接用于生成可靠的分类或本地级别的估计。州和国家一级的估算通常掩盖了地方一级的差异，这反过来又限制了与行政部门内部和部门之间在本地进行的扶贫政策的有效实施。本文使用NSSO的家庭消费者支出调查数据并将其与人口普查数据联系起来，得出印度西孟加拉邦农村地区贫困发生率的可靠的地区级估计。特别是，探索了小面积估算（SAE）方法以生成可靠的地区级贫困估算。结果清楚地表明，基于模型的SAE方法生成的地区级估计是准确且具有代表性的。还制作了一张地图，显示西孟加拉邦各州的贫困发生率如何变化。这项研究得出的估计数有助于满足不同国际组织以及印度政府各部委对政策研究和战略规划的数据要求。