Sections 2 to 4 have established where multidimensional poverty is concentrated, how it is generated and amplified, and who experiences it most acutely. This section turns to a final diagnostic question: what kind of poverty do these households experience? The MPI score decomposes into three mutually exclusive dimensional contributions (education, health and living standards) whose shares sum to 100 percent by construction; comparing these shares across contexts and household groups reveals shifts in the underlying structure of deprivation. Four findings emerge: health and living standards drive the national picture; dimensional composition varies by context, pointing to different intervention priorities; and similar MPI scores can mask fundamentally different poverty dynamics. Together, these findings indicate that contexts and groups with similar headline poverty levels often face very different compositional structures of deprivation, with direct implications for which interventions are likely to reduce poverty most effectively. 5.1 The National Picture: Health and Living Standards Dominate Nationally, Myanmar's MPI score of 0.150 is composed of deprivations in living standards (0.062, or 41.3% of the index), health (0.059, or 39.3%) and education (0.029, or 19.4%). Living standards captures deficits in housing quality, sanitation, water access and electricity; health captures access to healthcare and food security stress (measured through the Livelihood Coping Strategy Index - LhCSI); education captures schooling attendance and attainment. Together, the first two dimensions account for more than four fifths of national multidimensional poverty, indicating that the binding constraints on household well-being are basic living conditions and service access rather than education access itself. The urban-rural decomposition reinforces this finding. The rural MPI score (0.170) and urban MPI score (0.097) differ substantially in level, but the compositional shares are nearly identical: in rural areas, education accounts for 19.4 percent of the index, health 38.7 percent and living standards 42.0 percent; in urban areas, the corresponding shares are 19.7 percent, 42.1 percent and 38.2 percent. The urban-rural gap is a gap in overall poverty, not in its dimensional structure. This stable national composition serves as the reference against which variation across states/regions, household types and shock exposure is read in the sections below. 5.2 Variation Across States and Regions Dimensional composition varies across states and regions. Figure 11 shows both the MPI score (top panel) and the dimensional composition (bottom panel) for each state and region, sorted by MPI score. While MPI scores fall steadily from Rakhine (0.34) to Nay Pyi Taw (0.09), the dimensional composition does not shift in lockstep: each state has its own compositional profile reflecting its specific service environment, infrastructure conditions and schooling access. This cross-state variation in dimensional composition is significant, and it is not a product of MPI score itself. State-level MPI score is driven overwhelmingly by headcount, the share of the population that is multidimensionally poor, rather than by intensity, the average depth of deprivation among the poor: headcount ranges widely across states and regions, from 19.7 percent (Nay Pyi Taw) to 68.4 percent (Rakhine), while intensity varies only narrowly, from 44.2 percent (Nay Pyi Taw) to 50.1 percent (Rakhine) (Table 3). In other words, states and regions differ widely in how many people are poor, but the poor in any given state or region are, on average, similarly deep in poverty. Each 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 37