Poverty measurement (in India): Defining group-specific poverty lines or taking preferences into account?

We study absolute income poverty measurement when agents differ in preferences and face different prices. The difficulty arising from price heterogeneity is typically solved using equivalent income, but the choice of the reference price vector remains arbitrary. We provide a way to solve this arbitrariness problem by making the poverty measure consistent with preferences: an agent qualifies as poor if and only if she prefers the poverty line bundle to her current consumption bundle. We then prove that defining group/region specific poverty lines is another way of recovering consistency with preferences, provided one uses the headcount ratio. Comparing the resulting three approaches using Indian data, we find that the different approaches leads to different poverty conclusions. We show that not taking preferences into account leads to severely underestimating urban poverty.