An Analysis On The Formation And Efficiency Of Rural Risk-sharing Network

Early measurement of poverty regards poverty as lack of income or consumption,ignoring the impact of fluctuation of income and consumption on the welfare of thepoor. With the concept of vulnerability proposed by the World Bank being acceptedgradually, the academia learn more about the problem of poverty and the risk faced bythe poor, and the strategies to cope risk attracted the attention of the economists. Risksnot only directly affect the welfare of the poor, but also affect the behavior of theirproduction, and thus affect their long-term development. The risks faced by the poorliving in the backward areas of developing countries are more frequent and moresevere relative to other groups. Therefore if the shocks can’t be dealt with, the poorgetting out of poverty temporarily are likely to be knocked down again by unfavorableshocks and fall into vicious cycle of poverty trap.Because the problems of asymmetric information and enforcement of contractare very serious in the backward areas of developing countries, the formal insurancemarket is imperfect. A strong desire to survival prompts the poor to create a lot ofstrategies to deal with risks, however, including diversification of sources of income,sale of productive assets, etc. These strategies play a certain role in coping shocks,reflecting the living strategies of the poor. However these risk-coping strategies aremore or less defective. Some strategies even have to sacrifice long-term developmentfor short-term survival, which is tantamount to drink poison to quench thirsty.Among the risk-coping strategies of the poor, the most important and mostcommon one is risk-sharing, that is households share risks horizontally to defuse thenegative impact of shocks on individual household. Risk-sharing is similar to formalinsurance market except of lack of formal contract. Thus risk-sharing is thespontaneous formation of an informal institution arrangement.Empirical studies on informal risk sharing show that risk sharing often presentsthe form of network. In order to overcome the problem of asymmetric information,the informal risk-sharing is often limited to a network of relatives and friends. Basedon the empirical founding, we model the formation of risk-sharing network andsimulate it in the computer using Agent-based Computational Economics.This paper incorporates the formation and termination of risk-sharing relationsinto the framework of cost-benefit analysis, thus endogenize the formation of risk-sharing relations. Through simulation we find the risk-sharing network will nottend to stabilize, but in the process of evolving. Some new relations are formed, andsome old relations are terminated, which means the risk-sharing networks we observein real life maybe just a snapshot of the network in the process of evolution. Bychanging the corresponding parameters and comparing the networks formed at thesame point in time, we find that the completeness of the network is related to thedegree of risk aversion and income volatility. The higher the degree of risk aversion,the higher the degree of completeness of the network; the greater the degree ofincome volatility, the higher the degree of completeness of the network. Theseconclusions are consistent with our intuitions, also confirm that our model isreasonable.We also find that the agents will fall into different situation during the process ofevolution of the network. Some agents have more risk-sharing relation, while someothers have less. Some agents are in larger components, while some others in smalleragents. Some agents’ expected utilities are higher, while some others’ lower. Evenagents in the same component are in different situation. Agents on the edge ofcomponent have weaker risk-sharing relationship, thus are easy to be isolated. Takinginto account of the fact that all agents have the same initial conditions, we believe thephenomenon can’t be attributed to the difference of the agents. The difference insituations arises from the interaction among agents in the process of evolution, that is,the emergent property generated during the process of network formation. The findingmeans that the existing literature which explains the characteristic of the network maydraw biased conclusion.Another part of the empirical research on risk-sharing network focus on the testof its efficiency, research in this area basically refuses the assumption of fullefficiency. Theoretical research on the interpretation of this conclusion is generallyfocused on the risk-sharing behavior itself between internal relations, that is,attributing the inefficiency to the problems of adverse selection and moral hazard.This paper provides another explanation to the inefficiency of risk-sharingnetwork. We believe that the externality during the formation process of risk-sharingnetwork is the root cause of its inefficiency. By means of simulation we confirm thatonce the externality during the formation process of risk-sharing network isinternalized, complete network can be formed. In addition we introduce the concept offarsighted agent into the formation of risk-sharing network, and find that the more the farsighted agents, the higher the degree of he completeness of risk-sharing network.In addition, we introduce the problem of asymmetric information on the basis ofthe reference model, that is, agents possess complete information on their own incomebut incomplete information on others’ income. We find the lower the degree ofcompleteness of information, the lower the efficiency of the risk-sharing network. Bymeans of introducing the problem of asymmetric information to the formation ofrisk-sharing networks, this paper incorporates two explanations of inefficiency ofrisk-sharing network——the incompleteness of risk-sharing and the externality duringthe formation process of risk-sharing network——into one framework.