Social Networks, Financial Markets, and Household Organization in a Developing Economy

拟议的研究利用加纳 PI 收集的数据来研究以快速技术创新和土地资源紧张为特征的农村经济中的社会网络、金融市场和家庭组织。在两年的时间里，我们收集了关于投入、产出、土地权属和土壤化学的详细地块级数据、样本中个体之间的学习互动、资金流动和劳动力交易信息、地块级GIS数据以及关于经济活动（包括消费、非农收入和时间分配的面板数据）和社会经济背景的个体级数据。拟议的研究由三个主要部分组成。第一个是对社会学习和网络形成的调查。在调查地区，农业系统正在迅速从木薯-玉米间作转变为生产菠萝作为新鲜水果出口。这一转变涉及一系列新技术，包括对用于维持土壤肥力的休耕制度进行改造、利用新的产出市场，特别是密集使用以前农业制度中从未使用过的农用化学品。我提出了一种方法，用于识别从农民对这些新化学品进行的实验中获得的信息通过社交网络的传播情况。初步结果提供的证据表明，农民使用新化学品的水平对网络中直接或间接与该农民有联系的其他人使用这些化学品进行的实验结果敏感。鉴于这些结果，我建议对这些村庄的网络形成过程进行建模。要测试的假设是，一个人更有可能与可能拥有对该个人有价值的信息的其他人建立网络连接。其次，菠萝种植比种植替代作物更有利可图，但成本也更高。这就提出了采用过程受到研究区域金融市场特征影响的可能性。初步证据表明，财富的变化（视为内生的）是菠萝种植的重要决定因素。我建议使用资金流动数据以及支出、资产和时间使用面板数据来探索资本约束影响采用决策的假设。农民同时做出有关土壤肥力管理的决策。 由于这是一个跨期分配问题，因此它与每个人在资本市场中的立场密切相关。在非正式采访中，一些农民表示，他们正在降低土壤肥力以扩大生产。第二个目标是检查与土壤肥力有关的决策（使用土壤化学面板数据），并将其与收养决策和研究区域可用的金融市场联系起来。最后，加纳的证据表明，家庭内部资源分配机制既不符合单一家庭模型，也不符合更一般的集体家庭模型。特别是，家庭内部的风险分配不是帕累托有效的。有证据表明配偶之间存在重要的信息不对称。该研究计划的这方面将检验这样一个假设：家庭内部生产关系的内生结构是家庭内部信息流动以及家庭内部分配决策的决定因素。

The proposed research uses data collected by the PI in Ghana to examine social networks, financial markets and household organization in a rural economy characterized by rapid technological innovation and stress on land resources. Over a two year period, we collected detailed plot level data on inputs, outputs, land tenure and soil chemistry, information on learning interactions, financial flows and labor transactions between individuals in the sample, plot level GIS data, and individual level data on economic activities (including panel data on consumption, non-farm income and time allocation) and socioeconomic background.The proposed research consists of three major components. The first is an investigation of social learning and network formation. In the survey region, the farming system is rapidly changing from a cassava-maize intercrop to production of pineapples for export as fresh fruit. This shift involves a set of new technologies including a transformation of the fallow system that had been used to maintain soil fertility, the use of new output markets and, particularly, intensive use of agricultural chemicals where none had been used in the previous farming system. I propose a method for identifying the transmission through the social network of information derived from farmers' experiments with these new chemicals. Preliminary results provide evidence that the level of a farmer's use of the new chemicals is sensitive to the outcomes of the experiments with these chemicals by others to whom the farmer is directly or indirectly connected in the network. Given these results, I propose modeling the process of network formation in these villages. The hypothesis to be tested is that an individual is more likely to set up network connections with others who are likely to posses information valuable to that individual.Second, pineapple cultivation is both much more profitable and much more costly than the cultivation of alternative crops. This raises the possibility that the process of adoption is influenced by the characteristics of financial markets in the study area. Preliminary evidence is that variation in wealth (treated as endogenous) is an important determinant of pineapple cultivation. I propose to use the data on financial flows, along with the expenditure, assets and time use panel data to explore the hypothesis that capital constraints influence adoption decisions. Farmers simultaneously make decisions regarding the management of soil fertility. Since this is an intertemporal allocation problem it is intimately related to each individual's position vis-a-vis capital markets. In informal interviews, some farmers stated that they were drawing down soil fertility in order to expand production. The second goal is to examine decision making with respect to soil fertility (using the panel data on soil chemistry) and to relate this to adoption decisions and to the financial markets available in the study area.Finally, there is evidence in Ghana that the mechanism of intrahousehold resource allocation corresponds to neither the unitary household model nor the more general model of a collective household. In particular, the allocation of risk within households is not Pareto efficient. There is evidence of important information asymmetries between spouses. This aspect of the research program will examine the hypothesis that the endogenous structure of production relations within the household is a determinant of intrahousehold information flows and hence intrahousehold allocation decisions.