On the theory and measurement of relative poverty using durable ownership data

Poverty measurement using durable ownership data is an attempt to infer income constraints by observing consumption choices. But what drives household spending choices on durable goods? How do these choices relate to poverty and class? What does it mean to be 'relatively' poor and why should we care to measure it? In this paper, we propose an economic theory of household decision-making that links these questions using a novel wealth-begets-wealth mechanism. We show that the steady state distribution of total (accumulated) household durable expenditures in this model exhibits natural clusters (classes). Furthermore, certain households may be vulnerable to a long run 'poverty of opportunities', being unable to access any of the channels of income generation available in society. Our model shows that relative poverty can be understood as the endogenous outcome of an intergenerational process that perpetuates unequal access to opportunities. This finding has novel implications for the measurement of poverty, which has traditionally hinged on definitions that assume exogenous (often arbitrary) cutoffs. The contribution of this paper also lies in its novel methodology, viz., formulating a theoretical model as the foundation of a data-generating process for synthetic observations, using patterns observed therein to inform the process of poverty measurement. The methodology delivers a framework for generating testable hypotheses around the long-run effect of policy changes (such as income transfers or education subsidies) on relative poverty – an approach that can be applied generally to understand the observed behaviour of economic agents in complex dynamic settings.