Climate Variability, Migration and Intergenerational Transfers

描述（由申请人提供）：本项目借鉴并推进了我们对移民和家庭间转移作为墨西哥社区风险分担和消费平滑机制的理解。它特别研究了作为外源性环境风险来源的降雨变异性。长期年际变率衡量一个地点与农业相关的风险，而短期年际变率反映了一些自然随机和循环变率以及由于人为气候变化而增加的变率。因此，长期变异性的影响表明环境风险的影响已被编入社区中发展的社会制度，而短期变异性的影响则表明对环境风险的积极适应。相比之下，与长期平均降雨量的单年偏差的影响显示出对收入或生产冲击的直接反应。因此，该项目将区分将环境风险与人类行为联系起来的几种可能机制，为发展中国家家庭决策和人类适应气候变化的跨学科文献做出贡献。该项目使用可公开获得的社会调查数据——墨西哥家庭生活调查（MxFLS）——来衡量移民、家庭间转移和关键控制变量。然后，这些数据将与根据墨西哥256个气象站近35年来每小时降雨量数据构建的降水测量结果合并。这两个数据源将使用MxFLS数据中确定的最好的地理位置，以及与欧几里得距离最接近市政中心的气象站的记录相匹配的社会调查数据合并。永久和临时迁移的事件历史和logit回归模型将检验降雨的长期平均年际变率、短期年际变率和与长期平均降雨量的单年偏差的不同影响。家庭间转移的Logit回归模型检验了这些相同变量的影响。

DESCRIPTION (provided by applicant): This project draws on and advances our understanding of migration and inter-household transfers as risk-sharing and consumption-smoothing mechanisms in Mexican communities. It specifically examines rainfall variability as a source of exogenous environmental risk. Long-term interannual variability measures the risk associated with agriculture in a location, while short-term interannual variability reflects some natural random and cyclic variability and some increases in variability due to anthropogenic climate change. Thus, effects of long-term variability show the effects of environmental risk codified in social institutions that have evolved in a community while effects of short-term variability show active adaptation to environmental risk. Effects of single year deviations from the long-term average rainfall, in contrast, show immediate responses to income or production shocks. Thus, the project will distinguish between several possible mechanisms linking environmental risk to human behavior, contributing to the interdisciplinary literatures on household decision-making in developing countries and on human adaptation to changing climate. The project uses publicly available social survey data, the Mexican Family Life Survey (MxFLS), to measure migration, inter-household transfers, and key control variables. These data will then be merged with precipitation measures constructed from hourly rainfall data from approximately the last 35 years from 256 weather stations in Mexico. The two data sources will be merged using the municipio, the finest geography identified in the MxFLS data, with social survey data matched to the record from the weather station closest in Euclidean distance to the centroid of the municipio. Event history and logit regression models of permanent and temporary migration out of communities will examine the differing effects of long-term average interannual variability in rainfall, short-term interannual variability, and single year deviation from the long-term average rainfall. Logit regression models of inter-household transfers examine the effects of these same variables. PUBLIC HEALTH RELEVANCE: This project examines the ways in which climate variability influences human behavior, focusing on migration and resource-sharing within families. We analyze the relationships between changes in the amount of rainfall and the year-to-year variation in amount, and migration, and support of family members with money, time and goods, between 2001 and 2005 in communities across Mexico.