Collaborative Research: ESE: Food Price Spikes in a Warming World: Estimating Risks and Policy Responses

从2006年冬季到2008年夏季，世界上最重要的四种主食(大米、小麦、玉米和大豆)的价格上涨了两倍多。这四种主要作物约占世界卡路里基础的75%。该项目开发了一种新的方法，使用有关天气和农业产量的详细数据来研究气候变化如何影响粮食、商品价格，特别是商品价格的变异性。该项目首先使用计量经济学技术以及关于天气和农业产量的详细数据。这些数据被用来估计和检验天气和产量之间的非线性关系的统计模型；该模型的结果是对全球所有主要产区四种主要农作物的产量变异性的预测。下一步使用现有的气候变化预测，对未来天气模式的具体变化，包括平均/平均结果的变化，气候变异性的变化，以及不同生产区之间的相关性，制定合理的预测。然后，将前两个步骤的结果结合起来，预测气候变化将如何改变农业产量的分布。下一步是考虑过去的天气冲击如何影响这些主要作物的供需。这些信息被用来估计四种主要作物中每一种的供需结构模型。将这一市场模型与农业产量分布的预测变化相结合，使采购经理人能够预测未来气候变化将如何影响大宗商品价格。这些模型还将被用来检验政府政策变化以及农民和大宗商品交易商每年储存作物的方式变化可能产生的影响。研究计划的具体内容包括：将天气和产量变异性联系起来。PIs扩展了他们早期的工作，以检查全球四种主要农作物的产量变异性。B：气候变化预测。PIs汇总了各种大气环流模型对(I)平均结果；(Ii)变异性的变化；(Iii)生产区域之间的相关性的预测。天气模式的这三种变化中的每一种都会直接影响产量的分布。C：预测产量分布结合A和B部分给出了气候变化下的修正的产量分布。A部分中的非线性温度-产量关系表明，即使是恒定变化的平均气温上升也会影响产量变异性，天气变异性的变化也会影响产量变异性。最后，如果天气变得更多(或更少)相关，或者如果农业生产的地理位置变得更集中(或更集中)，特殊生产冲击将不再平均，世界总产量可能变得更多(或更少)变化。D：产量变异性和大宗商品价格。使用天气引发的产量冲击作为工具，研究人员可以估计四种主要作物的供需弹性。虽然随机产量冲击以前被用来估计需求弹性，但滞后产量冲击也可以用来确定供应弹性，因为过去的产量冲击与当前时期通过储存所做的努力有关。通过弹性估计，研究人员可以将他们的产量分布转化为大宗商品价格的分布。E：模拟竞争性储存和政府政策的影响。农民和大宗商品交易商可能会通过调整库存持有量来应对产量和价格波动的变化。面对年复一年的生产变化无常，库存有助于减弱价格变化性。因此，库存调整将在一定程度上缓冲生产变异性的变化，灌溉农业的持续扩张也是如此。另一方面，出口限制和其他政府政策可能会夸大价格的可变性。这个跨学科的项目将经济学家、农学家和环境科学家聚集在一起。研究结果将有助于指导政策制定者和农民适应未来的气候变化。

Between the winter of 2006 and the summer of 2008 prices of the world's four most importat staple food commodities (rice, wheat, corn and soybeans) more than tripled. These four staple crops comprise about 75% of the world's caloric base. This project develops a new methodology to use detailed data about weather and agricultural output to examine how climate change affects food commodity prices and particularly commodity price variability. The project begins with using econometric techniques and detailed data on weather and agricultural output. The data are used to estimate and test a statistical model of a non-linear relationship between weather and output; the results of that model are predictions about the yield variability of the four major staple crops in all major global production regions. The next step uses existing climate change projections to develop sensible predictions for future specific changes in weather patterns, including changes in the mean/average outcome, changes in the variability of climate, and correlation between the different production regions. The results of the first two steps are then combined to give a prediction of how climate change will change the distribution of agricultural yields. The next step is to consider how past weather shocks have affected supply and demand for these staple crops. This information is used to estimate a structural model of supply and demand for each of the four staple crops. Combining this market model with the predicted changes in the distribution of agricultural yields allows the PIs to make predictions about how future climate change will affect commodity prices. The models will also be used to examine the likely effects of changes in government policy and changes in how farmers and commodity traders store crops from year to year. The specifics of the research plan includeA: Linking weather and yield variability. The PIs extend their earlier work to examine yield variablility of the four major staple crops throughout the world.B: Climate change projections. The PIs summarize the predictions of various General Circulation Models on (i) mean outcomes; (ii) changes in variability, and (iii) correlation between production regions. Each one of these three changes in weather patterns would directly influence the distribution of yields.C: Predict yield distributionsCombining parts A and B gives revised yield distributions under climate change. The non linear temperature-yield relationship from part A implies that even a mean increase in temperatures with constant variance could impact yield variability, as would a change in weather variability. Finally, if weather becomes more (or less) correlated or if the geography of agricultural production becomes more (or less) concentrated, idiosyncratic productions shocks will no longer average out and total world production could become more (or less) variable.D: Yield variability and commodity prices.Using weather-induced yield shocks as instruments allows the researchers to estimate supply and demand elasticities for the four staple crops. While randome yield shocks have previously been used to estimate demand elasticities, lagged yield shocks can also be used to identify supply elasticities, as past production shocks are linked to the current period's effort through storage. With the elasticity estimates the researchers can translate their yield distribution into a distribution of commodity prices.E: Simulate the effects of competitive storage and government policies. Farmers and commodity traders are likely to resond to changes in production and price variability by adjusting inventory holdings. Inventories help to attenuate price variability in the face of year-t0-year productions variability. Inventory adjustments will therefore partly buffer changes in production variability, as could continued expansion of irrigated agriculuture. On the other hand, export restrictions and other government policies might exaggerate price variability.This interdisciplinary project brings economists, agronomists, and environmental scientists together. The results will be useful for guiding both policymakers and farmers as they adjust to future climate change.