Figure 11 Dimensional composition of MPI (by geography) (percentage contribution) Figure 11. Dimensional composition of MPI by state/region MPI scores vary widely across states/regions, while dimensional composition differs by local context Education MPI score Health Contribution to MPI (%) Living standard Rakhine 0.342 12% 42% Kayah 0.328 9% 49% 42% 12% 43% 46% Chin 0.220 Sagaing 0.196 Kachin 0.188 16% 0.183 15% Ayeyarwady Kayin 0.153 Mon 0.149 Mandalay 0.148 46% 23% 37% 40% 37% 22% 40% 43% 48% 43% 35% 28% 34% 38% 25% 38% 36% Tanintharyi 0.140 18% 43% 39% Shan 0.131 20% 40% 40% Magway 0.128 19% 42% 39% Bago 0.118 21% 38% 42% 19% 41% 41% Yangon 0.092 Nay Pyi Taw 0.087 25% 42% 33% National: 0.150 Source: PPS 2025 (UNDP) state's compositional profile, shown in Figure 11, therefore reflects contextual differences in service infrastructure, conflict exposure, livelihood patterns and access, rather than tracking each state's position in the MPI severity ranking. Some states/regions, such as Rakhine, Kayah, Chin, Kachin and Ayeyarwady, exhibit compositions where health and living standards together dominate, with education compressed. Others, such as Sagaing, Mon, Mandalay and Nay Pyi Taw, show more balanced contributions across all three dimensions. Still others fall between these patterns. These compositional differences, determined by geography rather than MPI score, establish the foundation for determining which interventions are likely to carry the greatest leverage in each setting. However, it should be noted that a lower education share in a given state/region does not imply an absence of education deficits; rather, it indicates that health and living standards account for a larger proportion of each household's total deprivation, making service restoration and infrastructure investment the most efficient entry points for poverty reduction in those settings. Health is the most stable dimension across all states/regions, ranging from 34 percent (Mon) to 49 percent (Kayah) without a consistent relationship to overall MPI level. 5.3 The Household-Level Gradient: Composition Shifts as Deprivation Deepens Section 5.2 shows that dimensional composition across states and regions does not track state-level MPI severity, since MPI score is driven by headcount rather than intensity. This section turns to a different question: not where households are, but how the structure of their deprivation changes as poverty deepens. 38 LAYERS OF DEP RIVATION : Mu ltid im e n s io n a l Pov er t y i n M y a n m a r