讲座题目：Segregation, Spillovers, and the Locus of Racial Change（隔离、溢出效应与种族变迁的空间轨迹）

主讲人：Donald Davis 哥伦比亚大学经济学与国际事务讲席教授

讲座时间：2025年04月28日15:30

讲座地点：学院440

讲座内容摘要：

Existing empirical research in economics on neighborhood racial sorting is overwhelmingly premised on the idea that racial preferences for a location depend on the racial shares in that location, without considering potential spatial spillover effects from nearby areas. Does this matter for the way we view the cross-section and dynamics of racial neighborhood segregation? We nest Schelling (1971)’s bounded neighborhood and spatial proximity theories within a discrete choice model, where the key distinction is precisely such spatial spillovers. We simulate the model and examine the data for 1970-2000 for more than 100 U.S. metros. Two features of the data are most compelling: the powerful presence of racial clusters and the fact that drastic racial change is concentrated at the boundary of these clusters. Both point to the spatial proximity model as the proper foundation for a theory of racial neighborhood evolution. We use these insights to revisit prominent results on racial tipping where our theory guides us to distinguish differences by location. While prior research pointed to powerful racial tipping in the form of White exit, we show this is largely driven by theoretically-distinct “biased white suburbanization” leading to White entry in remote areas. In urban areas far from existing Minority clusters, we find zero or small tipping effects, at odds with a bounded neighborhood interpretation. The most consistent effects of tipping, still of modest size, are found in areas adjacent to existing Minority clusters, confirming the relevance of the racial spillovers of the spatial proximity model. Existing research conflates these quite distinct effects. Overall, our results suggests that tipping is a less central feature of racial neighborhood change than suggested in prior research and that greater attention needs to be paid to spatial dimensions of the problem.

现有经济学文献关于居住区种族分选的实证研究大多数基于一个核心假设：个体对居住地的种族偏好仅取决于该区域本身的种族构成，而忽视了邻近地区可能存在的空间溢出效应。这种理论预设是否会影响我们对种族居住隔离的截面特征与动态演变的理解？本研究将谢林（1971）提出的社区边界理论与空间邻近理论嵌套于一个离散选择模型中，其核心区别正是这种空间溢出效应的存在。通过模型模拟与对1970-2000年间100多个美国大都市区数据的实证检验，我们发现两个最具说服力的特征事实：显著的种族集聚现象，以及显著的种族变化集中于这些集聚区边界。二者共同指向空间邻近模型可以作为解释种族居住区演变的理论基础。基于此理论视角，我们重新审视了关于种族“临界点”（tipping）的经典结论。既有研究强调以白人迁出为主导的显著的种族临界点效应，但我们证实这种现象主要由理论上迥异的“偏向性的白人郊区化”致的——即白人向偏远地区迁入。在远离现有少数族裔聚集区的城市区域，我们观测到的临界点效应为零或微乎其微，这与社区边界理论的解释形成显著差异。最一致的发现是有较弱的临界点效应出现在邻近少数族裔聚集区的边缘地带，印证了空间邻近模型中种族溢出效应的解释力。既有研究未能有效区分这些效应的本质差异。总体而言，我们的研究表明：既有文献强调的临界点现象在种族聚居区变迁中并非核心机制，而空间维度亟需得到更多关注。

主讲人学术简介：

Donald Davis，现任哥伦比亚大学经济学与国际事务讲席教授。他于1992年获得哥伦比亚大学经济学博士学位，1992至1999年间任教于哈佛大学，曾任哥伦比亚大学经济系主任，并担任American Economic Review、Journal of International Economics、Journal of Urban Economics等多家学术期刊编委或主编，城市经济学会前任主席。Davis的研究涵盖空间经济学的多个领域，从国际贸易到经济地理，从区域经济学到城市经济学均有涉猎。其学术成果包括：贸易理论检验、国际贸易中的市场规模优势、贸易与不平等问题研究，以及近期关于不同规模城市间技能分布差异、城市内部消费的种族与族群隔离等主题。论文发表于 American Economic Review (8篇)、Journal of Political Economy、Review of Economic Studies等经济学顶刊以及 Journal of International Economics、Journal of Urban Economics等国际贸易、城市经济学领域顶刊。