Collaborative Research: Trade Costs, Preferences, and Government Policies: Understanding Market Outcomes in the World Automobile Industry

该项目的目标是量化市场细分的来源，并了解贸易和投资摩擦对世界汽车业的影响。该框架使采购经理人能够评估贸易成本、对国内品牌的偏好和需求异质性的重要性。物价指数还量化了关税和非关税壁垒(例如，由于燃料价格和环境监管的差异而产生)对市场结果的影响。该项目将是第一个利用几大洲的数据对汽车需求和边际成本进行一致估计的项目。这种地理差异使采购经理人能够估计各种与贸易有关的成本，同时考虑到模型水平的未观察到的异质性和本土偏见。这种方法将政策设置的壁垒与运输成本等技术壁垒分开。他们的分析范围将促进对贸易成本、偏好异质性和政府政策之间相互作用的了解，这些政策在塑造汽车行业和更广泛的国际贸易中的市场结果。这项研究将评估贸易成本与偏好差异在世界汽车贸易中的重要性-汽车是一个主要的制造业。这将导致对公司进行外国直接投资和进行特定市场模式开发的动机的进一步了解。包含装配地点的数据集将是分析FDI在汽车行业的重要性的宝贵资源。对汽车行业关税和非关税壁垒重要性的估计将为美国和欧盟正在进行的贸易和投资伙伴关系谈判的政策制定者提供信息。为了进行这项研究，PI首先构建一个新的数据集，将咨询公司Polk Automotive收集的来自三个不同大陆的九个国家的销售数据与WardsAuto收集的按车型进行的世界组装地点普查联系起来。他们链接这些数据集，以产生关于销售、价格、特性、组装和总部位置的模型级数据。其次，他们使用这些数据来估计国际汽车需求的结构模型。他们扩展了Berry，Levinsohn和Pakes(1995)首创的技术，提出了一个模型，该模型考虑了国家之间和国家内部对特性和收入的喜好的异质性，以灵活地估计每个国家汽车的弹性。这个需求模型将衡量家庭偏见的重要性。或者消费者倾向于更喜欢本国制造商，而不是外国制造商。他们模型的一个关键特征是将这些需求驱动的解释与供应方驱动因素分开。第三，他们使用价格和需求弹性来构建向每个市场供应每辆汽车的边际成本估计，并使用这些估计来估计模型的供应方。展望未来，成本和需求估计可以用来分析公司吗？位置和营销决策，并研究潜在政策干预的影响。

The goal of this project is to quantify the sources of market segmentation, and to understand the implications of trade and investment frictions on the world automotive industry. The framework enables the PIs to assess the importance of trade costs, preferences for domestic brands, and demand heterogeneity. The PIs also quantify the impact of tariff and non-tariff barriers (e.g. arising from differences in fuel prices and environmental regulation) on market outcomes. The project will be the first to develop consistent estimates of automobile demand and marginal costs using data across several continents. This geographical variation allows the PIs to estimate a variety of trade-related costs while accounting for model-level unobserved heterogeneity and home bias. This approach disentangles barriers set by policies from technological barriers such as transport costs. The scope of their analysis will advance the state of knowledge on the interactions between trade costs, preference heterogeneity, and government policies in shaping market outcomes in the automobile industry and in international trade more broadly. The study will evaluate the importance of trade costs versus preference differences in the world trade in automobiles---a major manufacturing industry. This will lead to further insights into understanding the incentives of firms to conduct foreign direct investment and conduct market-specific model development. The dataset with assembly location included will be a valuable resource for analyzing the importance of FDI in the auto industry. The estimates on the importance of tariff and non-tariff barriers in the automotive industry will inform policy makers for the ongoing Trade and Investment Partnership negotiations between the U.S. and the EU. To undertake this research, the PIs first construct a new data set by linking sales data from nine countries on three separate continents, collected by the consulting firm Polk Automotive, with a world census of assembly locations by model---collected by WardsAuto. They link these datasets to yield a model-level data on sales, prices, characteristics, assembly and headquarters location. Second, they use this data to estimate a structural model of international demand for automobiles. Extending techniques pioneered by Berry, Levinsohn, and Pakes (1995), they propose a model that accounts for inter and intra country heterogeneity in tastes for characteristics and income to flexibly estimate elasticities for automobiles in each country. This demand model will measure the importance of ?home bias? or the tendency of consumers to prefer their local manufacturers to foreign manufacturers. A key feature of their model is to separate these demand driven explanations from supply side drivers. Third, they use prices and demand elasticities to construct estimates of the marginal cost of supplying each automobile to each market, and use them to estimate the supply side of the model. Going forward, the cost and demand estimates can be used to analyze firms? location and marketing decisions and study the impact of potential policy interventions.