01 1. Differentiated geographic targeting: Severity and scale require different responses The analysis identifies a systematic tension in geographic prioritization between severity and scale that cannot be resolved by ranking states/regions on a single indicator. In Kayah and Rakhine, poverty is exceptionally severe: MPI headcounts exceed 68 percent, and poor households are deprived in roughly half of all indicators simultaneously. These are the clearest cases of context-dominant poverty, where conflict, displacement, mobility restrictions and fragmented services push deprivation far beyond what household characteristics alone would predict. In these settings, area-based responses that restore essential services, basic living conditions and local markets are the primary entry point; household-level targeting alone will not address the contextual pressures driving the gap between observed and predicted poverty. In Rakhine specifically, responses must additionally account for the situation of stateless and displaced populations, including barriers related to documentation, freedom of movement and service entitlement, which standard area-based programming is unlikely to reach without explicit design adjustments. As noted in Section 1.1, the MPI estimates for Rakhine almost certainly understate true deprivation given these populations' limited representation in the survey frame, meaning the scale of need is greater than the figures in this report reflect. The scale of need in Rakhine is almost certainly greater than the figures in this report reflect. Stateless and displaced populations, including those facing barriers related to documentation and freedom of movement, fall outside standard survey frames, making them the hardest to count and the hardest to reach. The largest absolute concentrations of MPI-poor people, however, are not in Kayah or Rakhine. Ayeyarwady (2.65 million MPI-poor), Sagaing (2.38 million), Mandalay (2.11 million) and Shan (1.97 million) together account for roughly half of the 17.9 million poor nationally, despite recording only moderate MPI scores. In these regions, the scale of poverty reflects large populations with widespread but moderate deprivation. Scale-oriented approaches, including broad-based rural development, livelihood expansion and service coverage, are appropriate here alongside targeted household-level support. Township-level evidence reinforces the importance of within-state geographic differentiation. Poverty clustering within states/regions means that state-level averages conceal townships with substantially higher deprivation: in Ayeyarwady, for example, Mawlamyinegyun and Wakema exceed 70 percent MPI headcount, a severity comparable to the worst-affected townships in Rakhine and Kayah, despite the region's moderate state-level score. Geographic targeting should therefore operate at sub-state level, not solely on state/region rankings. 022. Area-based recovery or household investment: Let the driver decide Multidimensional poverty in Myanmar arises from two distinct but interacting sources, namely structural household disadvantages and contextual area-level conditions, and the relative weight of each varies considerably across states/regions. Identifying which driver dominates in a given context is not an academic distinction: it determines whether area-based or household-level responses are likely to produce the larger reduction in poverty. L AYE R S O F DE P R I VAT I O N : M u l t i di m en si on a l Pov er t y i n M y anmar 43