Econometric Demand and Welfare Analysis

政府的市场干预在现实生活中无处不在，包括对产品和服务的税收和补贴，对行业的监管和对市场力量的调节，扩大学校选择的集水区，引导信息不足的决策者做出最佳保险产品选择的行为推动和补贴，等等。拟议的项目旨在开发新的计量经济学工具，利用微观层面的数据对这些干预措施进行实际的成本效益分析。这种计算的关键挑战是通过消费者福利来衡量利益，这等于政策干预带来的变化对消费者的货币价值，这通常是研究人员无法观察到的。此外，当消费者的偏好不同时，收益和成本是异质的，因此需要统计分析来表征整个人口的福利分配。拟议的项目旨在开发计量经济学工具来实现这种分析。这些工具将有助于(1)从观察到的选择行为中理解消费者偏好的本质，以及(2)应用这些知识来评估现有和潜在的公共政策在各种重要的现实生活环境（如教育、住房、医疗保健、通勤旅行和度假住宿）中对社会福利的影响。该建议由四个部分组成，解决以下具体问题：(A)如何将社会总福利纳入计量经济计划评估和政策的最佳目标；现有文献通常关注个体结果作为其研究对象，而不是特定结果对个体本身的价值。(B)如何在享乐环境中衡量消费者福利，在这种环境中有许多可能的选择，每个选择都有几个个人偏好的特征，比如住房选择中的当地便利设施和犯罪，择校中的师生比例和学业成绩。(C)如何测试个人选择保险计划的最优性和保险市场中助推者的福利潜力，使用选择保险和随后索赔的微观数据。(D)如何利用从在线房源中收集的“大数据”来分析传统和在线供应商市场的需求模式和消费者福利，并应用于计算共享住宿服务Airbnb对福利的影响。这些子项目的共同主题是发展新的计量经济学方法，用于分析基于严格经济理论的实际福利计算的横截面数据。最终目标是通过对研究人员未观察到的个体偏好的异质性施加最小或不施加假设来获得可信的实证结果。这与现有计量经济学文献中可用的福利分析经验工具形成对比，这些工具依赖于对消费者偏好及其异质性的分析方便但经济上未经证实的临时假设。因此，拟议的项目将建立在由PI启动的研究计划的基础上，该计划由欧洲研究理事会在2016-21年期间提供为期5年的整合者资助，并将其范围从实质上扩展到以前未探索的背景，并在技术上分析新类型的数据，以便为与这些背景相关的基本政策问题提供可靠的答案。

Market-interventions by the government are ubiquitous in real life and include taxes and subsidies on products and services, regulation of industries and moderation of market-power, expansion of catchment areas for school-choice, behavioral nudges and subsidies to steer under-informed decision-makers towards their optimal choice of insurance products, and so forth. The proposed project aims to develop novel econometric tools for practical cost-benefit analysis of such interventions using micro-level data. The key challenge of such calculations is the measurement of the benefits via consumer welfare, which equals the monetary value to the consumer of the change resulting from the policy intervention, that is typically unobserved by a researcher. Further, when consumers differ in preferences, the benefits and costs are heterogeneous, hence statistical analysis is required to characterize the distribution of welfare across the population. The proposed project aims to develop econometric tools to enable such analysis. These tools will be useful for (1) understanding the nature of consumer preferences from observed choice behavior, and (2) for applying that knowledge to evaluate existing and potential public policies in terms of their impact on social welfare in a variety of important real-life contexts such as education, housing, healthcare, commuter travel and holiday accommodation. The proposal consists of four parts addressing the following specific questions:(A) How to incorporate aggregate social welfare in econometric program evaluation and the optimal targeting of policies; the existing literature typically foucuses on an individual outcome as its object of study, as opposed to how much that particular outcome is worth to the individual herself.(B) how to measure consumer welfare in hedonic settings where there are many possible alternatives, each characterized by a few features that individuals have preference over, such as local amenities and crime in housing choice and student-teacher ratio and academic outcomes in school-choice.(C) how to test optimality of individuals' choice of insurance-plans and the welfare potential of nudges in insurance markets, using micro-data on choice of cover and subsequent claims.(D) how to use `big data' collected from online listings to analyze demand patterns and consumer welfare in markets with both traditional and online providers, with an application to calculating the welfare impact of the shared accommodation service Airbnb. The common theme of these sub-projects is the development of novel econometric methods for analyzing cross-sectional data for practical welfare calculations that are founded on rigorous economic theory. The ultimate goal is to obtain credible empirical results by imposing minimal or no assumptions on heterogeneity in individual preferences unobserved by researchers. This is in contrast to available empirical tools for welfare analysis in the existing econometric literature which rely on analytically convenient but economically unsubstantiated ad-hoc assumptions on consumer preferences and their heterogeneity. As such, the proposed project will build on the research programme started by the PI that was funded by a 5-year Consolidator grant by the European Research Council between 2016-21, and expand its boundaries both substantively to previously unexplored contexts and technically to analyzing new types of data in order to provide credible answers to fundamental policy questions relevant to these contexts.