| Summary: | Production and exchange systems in agrarian subsistence economies have been described as following a "safety-first" decision-making principle that minimizes the risk of shortfall. Here, we extend and formalize this idea, with a particular focus on hunter-gatherer economies. Most hunter-gatherers are exposed to substantial livelihood risk because of the variable nature of their resource base. The primary mechanism of risk management observed among hunter-gatherers is food-sharing. We ask what kinds of network structures are most effective for managing the livelihood risks faced by hunter-gatherers, recognizing that such networks are epiphenomena of interactions between actors trying to minimize individual (or family) risk. This implies a trade-off between structures that effectively spread risk versus those that prevent disturbances from propagating. Echoing the classical anthropological literature, we suggest that generalized exchange is the foundation for subsistence risk-management networks. While the micro-foundations of generalized exchange have historically been couched in terms of cyclical tie-formation, we show that networks formed by a preponderance of cyclical relations would not satisfy the fundamental risk-management objective. Instead, we suggest that transitive closure, a process typically associated with hierarchical social relations, should be the primary mechanism of sharing-network formation. Networks formed by strong transitive closure are characterized by high tie density that produces many paths over which generalized exchange can be realized, a process that leads the resulting networks to be characterized by a strong core-periphery structure. We test this hypothesis with data on country-food sharing among Inuit hunters in Nunavik, finding that the food-sharing network has very high transitivity and is characterized by a striking core-periphery structure. We conclude with a discussion of network structure under different modes of livelihood risk management and what these networks tell us ...
|
|---|