Informal Risk-Sharing and Poverty Persistence

An old intuition suggests that poverty may perpetuate itself due to mild risk taking. Risk-averse individuals remain in poverty while their more daring peers escape by means of high-return, high-risk activities. As a growing literature increasingly highlights the role of insurance as a trigger for income-enhancing choices, several implications arise — informal risk-sharing agreements have found new attention, private microfinance and micro-insurance products have been promoted, and also public funds have been called to provide the poor with safety nets and thus encourage them into more entrepreneurial choices. However, the evidence thus far has consistently pointed out the limits of these insurance mechanisms. In the case of informal risk-sharing, its inability to cope with village-level shocks is well recognised. Furthermore, these arrangements may fail if partakers are not sufficiently similar in at least a few certain characteristics, i.e. each individual may need to bear in mind that certain choices are banned for her, unless she is willing to relinquish the protection provided by the arrangement. In this sense, when the activity promising enough earnings to escape poverty is at odds with those network requirements, the risk-sharing arrangement may become a poverty trap. The model in this paper depicts an instance of this scenario. Here, some members take the risk and leave, while others stay behind. The ability of the network to retain at least some of its members (and trap them into the low-earnings choice) will depend crucially on its initial size and on the initial distribution of income (human capital) among partakers. The model thus sheds light on the conditions under which the network collapses. At any rate, the literature has already made a cogent case for public resources to be devoted to the promotion of insurance products, either directly or through market-driven mechanisms. The paper is in keeping with this view. A planner intending to attract workers into better-paid (but say, distant, lonely) jobs will need to look into the safety nets potential employees will be hoping for. However, the model also suggests that, in some cases, a budget-constrained government should refrain from intervening — one by one, individuals will drift away from the network and migrate to the higher-earnings activity. Thus, public funds should not be necessarily allocated to promote the highest-earnings activities, but to those competing with a too big, too strong risk-sharing network. While the model accommodates other interpretations, it can be useful to think initially of the choice between staying in the friendly rural village and venturing out into an anonymous, lonely life in the big city, where earnings are however expected to be significantly higher. Some individuals will then dread life away from their home village, since they know no peer support will come to the rescue if misfortune strikes. Importantly, individuals are all identical, except for their human capital, which is exogenous to them, time-invariant, and observable to all others in the network. In particular, they are all equally risk-averse. Thus, those who remain in poverty are not too fearful, but just as risk-averse as the rich. Should a poverty trap exist, its rationale will lie elsewhere, namely on the distribution of human capital endowments and the initial network size. The existence of a poverty trap will depend on large enough networks and, also, on initial distributions where the poverty headcount is too high. In this sense, our result is at odds with the attention which has been paid to the advantages derived from social cohesion and reciprocal trust. Section 2 surveys the literature on a number of issues closely related to our main arguments. Section 3 presents the model and its implications, and section 4 provides a particular example as an illustration. Section 5 concludes.