“…While there is no conclusive explanation for this, it is likely due to an underestimation of consumption levels (Morduch, 2004;Townsend, 1994). As a consequence, the magnitude of the regression constant of household PLU estimates is consistent with widespread consumption poverty headcounts of 76% for VLS1 and 22% for VLS2 (Badiani et al, 2007). Second, the period over which households have been observed is long.…”