Non-linear relationship between income and multidimensional poverty15 Figure 8 Figure 8. Non-linear relationship between income and multidimensional poverty Predicted probability of being MPI-poor by monthly per-capita income (MMK, log scale) Predicted probability Income axis shown on a log scale Structural trap Predicted probability of being MPI-poor Initial income gains yield limited poverty reduction Transformative middle Diminishing returns Largest marginal reductions in poverty risk Remaining deprivation is increasingly structural 65% 19% 5% 30,000 100,000 300,000 1,000,000 Monthly per-capita income (MMK, log scale) Model: penalized logistic regression with natural cubic splines (df = 3), survey-weighted Source: PPS 2025 (UNDP) Labour insecurity as a transmission channel Employment type reinforces income vulnerability. Households relying on casual labour record an MPI headcount of 48.5 percent and a score of 0.228, among the highest across all employment groups. Figure 9 illustrates the gradient across employment types. The mechanism is not simply low wages: casual work provides no employment protection, no income continuity and no buffers against disruption. When livelihoods are interrupted by illness, conflict or natural hazards, casual workers have no savings or formal safety nets to draw on. Labour insecurity is therefore a transmission channel through which income shocks cascade into multidimensional deprivation. By contrast, households headed by employees show markedly lower levels of multidimensional poverty (MPI headcount 21.9%, MPI score 0.098). Employer-headed households, defined as those whose head runs a business with at least one hired worker, record the lowest poverty of any employment group (MPI headcount 20.4%, MPI score 0.085) and the lowest intensity (41.7%). The poverty gradient across employment types runs from employer to employee to own-account to casual worker, consistent with the interpretation that capital accumulation and the capacity to hire labour are stronger poverty buffers than wage employment alone. Own-account workers occupy a distinct intermediate position (MPI headcount 28.1%, MPI score 0.127): better protected than casual workers but meaningfully more vulnerable than either employees or employers, reflecting the absence of employment security and the thin buffers typical of self-employment in Myanmar's informal economy. 15 The x-axis uses a log scale to reflect the right-skewed distribution of monthly per capita income in Myanmar: most households are clustered at low incomes with a long upper tail. The log scale spreads out the lower range where most households are concentrated, making the steep poverty decline more clearly visible. The regression model is fitted on log-transformed income using a natural cubic spline (three degrees of freedom) with Firth penalization to capture the non-linear relationship between income and MPI poverty risk. Income is measured in Myanmar Kyat (MMK) per capita per month. 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 31