In traditional policy discourse, poverty is framed as an annual headcount, computed as the share of households, whose welfare, measured by expenditure or income, falls below the poverty line in a given year. Most low and middle income countries, including India, measure this using consumption, while income is often used as a proxy where consumption data is unavailable. While such measures capture who is poor, they fail to provide meaningful insight into the diverse ways in which poverty is experienced. This limitation arises from the implicit assumption that households can perfectly and equally smooth their consumption year-round, an assumption that rarely holds in practice. As a result, annual measures classify households as either always poor or never poor, obscuring substantial intra-year and inter-regional variability in the poverty experience. This paper argues that a nuanced understanding of poverty demands an inquiry that goes beyond ‘who’ is poor and looks at ‘when’ they are poor. Building on the framework of ‘Poverty at Higher Frequency’ introduced by Merfeld and Morduch (M&M 2022, 2025), I use household-level data from the Consumer Pyramids Household Survey (CPHS) to show that: (i) poverty is more widespread than suggested by yearly headcount measures, (ii) households’ ability to smooth consumption varies with socioeconomic status and directly shapes their poverty experience, and (iii) transitions in and out of poverty are not as clear-cut as implied by annual measures. I extend the framework by constructing a dynamic, region and state specific monthly poverty line. It inflates the 2011-12 poverty lines, estimated by the Planning Commission (2013) using the Tendulkar methodology (2009), with the state-level Consumer Price Index published by the Ministry of Statistics and Programme Implementation (MoSPI, 2024). The findings, drawn from recent, nationally representative data, are consistent with those of M&M, and show that the divergence between monthly and annual poverty measures is deeply heterogeneous and consequential for the design of anti-poverty interventions.
Read the full paper here
