中文摘要
美国各地区就业率差异长期存在。原则上，人口流动应当消除这些差异。移民能否发挥这一作用？值得注意的是，我发现自1960年以来，来自国外的新移民占针对这些差异的平均人口响应的40%，这远高于其在历史迁移总流量中的占比。但尽管如此，移民并未显著加快本地人口调整（或缩小本地就业率差异），因为它挤出了国内人口流动的贡献。事实上，这种挤出效应有助于解释国内人口流动率同期下降的现象。最后，我将国外移民的这种“超额”贡献归因于一种本地滚雪球效应，该效应由持续性的本地冲击和移民聚居区的动态所驱动。这一机制对将移民聚居区普遍用作外来流入工具变量的做法提出了挑战。但我并未因此放弃该工具变量，而是提出了一种（受本文模型启发）的实证策略来克服这些挑战，并验证了其有效性。
Abstract
The US suffers from persistent regional disparities in employment rates. In principle, these disparities should be eliminated by population mobility. Can immigration fulfill this role? Remarkably, since 1960, I show that new migrants from abroad account for 40% of the average population response to these disparities - which vastly exceeds their historic share of gross migratory flows. But despite this, immigration does not significantly accelerate local population adjustment (or reduce local employment rate disparities), as it crowds out the contribution from internal mobility. Indeed, this crowd-out can help account for the concurrent decline in internal mobility. Finally, I attribute the “excess” foreign contribution to a local snowballing effect, driven by persistent local shocks and the dynamics of migrant enclaves. This mechanism raises challenges to the (pervasive) application of migrant enclaves as an instrument for foreign inflows. But rather than abandoning the instrument, I offer an empirical strategy (motivated by my model) to overcome these challenges; and I demonstrate its efficacy.