Doctoral Dissertation Research in Economics: Econometric Evaluation of Costa Rica's Payments for Environmental Services: Scaling up Sample Size and Study Area

在经历了二十世纪后半叶世界上森林砍伐率最高的国家之一之后，哥斯达黎加在过去十年中获得了森林覆盖率。与此同时，哥斯达黎加政府实施了环境服务付费计划，该计划旨在鼓励私人土地所有者进行保护，他们拥有哥斯达黎加一半以上的森林。虽然这是为环境服务直接付费的最著名的例子之一，但明显缺乏经验证据来确定该项目是否以及在多大程度上促进了哥斯达黎加森林覆盖率的增加。该项目将(1)详细了解哥斯达黎加环境服务支付方案实施过程中的时间和空间变化，以及影响不同规模(家庭和地区)参与的因素；(2)通过扩大样本规模，在一个案例研究区域扩展由国家科学基金资助的对该方案的持续评估；以及(3)在区域范围内提供“概念证明”匹配。这些假设是：(1)保护费用产生森林面积的净增加；(2)使用家庭层面的数据对哥斯达黎加项目对森林覆盖的因果影响的估计与使用地区层面的次要数据获得的估计值相当。拟议的论文将建立在现有的国家科学基金会资助的研究(SES，0519194)的基础上，题为《公共产品的私人提供：应用匹配的估计器来评估热带森林保护的直接支付》。收集了关于家庭和财产一级的社会经济和生物物理特征的横断面数据，收集了方案参与者和非参与者的样本，并与森林复盖变化的分类遥感图像相结合。拟议在2007年上半年进行的实地考察工作将通过以下方式补充现有数据：(1)对主要参与者进行半结构化访谈，以了解1997-2005年该计划实施过程中的时空变化；(2)使用新的图像来源对森林覆盖进行分类；(3)汇编地理信息系统层，包括生物物理和社会经济特征以及哥斯达黎加所有人口普查段的边界；(4)开发关于土地机会成本的地理参考信息。广泛影响：拟议的研究建立在国家科学基金资助的一项试点研究的基础上，该研究旨在将项目评估技术应用于广泛推广的保护政策。通过对哥斯达黎加项目的大样本和全国范围的评估，这项研究为热带雨林保护政策工具的辩论提供了信息。这项研究还通过纳入空间信息、测试不同的分析尺度和估计发展中国家保护政策的因果影响，为项目评估的文献做出了贡献。拟议的实地工作产生了一个案例研究区域和整个哥斯达黎加的丰富数据集。拟议的分析结果包括会议报告、向同行评议期刊提交的报告、对哥斯达黎加学生的培训以及一篇博士论文。

After experiencing one of the world's highest rates of deforestation in the latter half of the twentieth century, Costa Rica has gained forest cover in the past decade. At the same time, the Costa Rican government implemented the Program of Payments for Environmental Services, which is designed to encourage conservation by private landholders, who own more than half of Costa Rica's forest. Although this is one of the best known examples of direct payments for environmental services, empirical evidence to determine whether and to what degree the program contributed to the increase in forest cover in Costa Rica is notably lacking. This project will (1) develop a detailed understanding of the temporal and spatial variation in the implementation of the Costa Rican Program of Payments for Environmental Services and of the factors that influence participation at different scales (household and regions) (2) extend an on-going NSFfunded evaluation of this program in a case study region by scaling up the sample size, and (3) provide 'proof of concept' matching at a regional scale. The hypotheses are that (1) conservation payments generate a net increase in forest area; (2) estimates of the causal effects of the Costa Rican program on forest cover using data at the household level are comparable with estimates obtained using secondary data at the regional level.The proposed dissertation will build on existing National Science Foundation funded research (SES, 0519194) entitled Private Provision of Public Goods: Applying Matching Estimators to Evaluate 'Direct Payments' for Tropical Forest Conservation. Cross sectional data on socio-economic and biophysical characteristics at the household and property level have been collected for samples of participants and nonparticipants in the program and combined with classified remote sensing imagery of forest cover change. The proposed fieldwork to be undertaken during the first semester of 2007 will complement existing data through (1) semi-structured interviews with key players in order to understand the spatial and temporal variation in implementation of the program from 1997 to 2005; (2) classifying forest cover using new sources of imagery; (3) compiling GIS layers including biophysical and socio-economic characteristics andboundaries of census segments for all of Costa Rica, and (4) developing geo-referenced information on land opportunity costs.Broader Impacts: The proposed research builds on a pilot study already funded by NSF to apply program evaluation techniques to a widely-promoted conservation policy. Through large-sample and national-scale evaluation of the Costa Rica program, this study informs the debate over policy tools for conservation of tropical rainforest. This study also contributes to the literature on program evaluation by incorporatingspatial information, testing different scales of analysis, and estimating causal effects of conservation policy in a developing country. The proposed fieldwork results in a rich dataset for both a case study region and all of Costa Rica. The proposed analyses results in conference presentations, submissions to peer reviewed journals, training for Costa Rica students, and a doctoral dissertation.