The Great, Greater, and Greatest Recessions of US States

This paper examines state-level differences in the timing, depth, and total employment effects of the Great Recession. It finds that several states were in recession prior to the official start of the recession, while more than a dozen states didn’t enter recession until six months or more after it. States’ exits from recession were similarly staggered. As a result, 11 states’ recessions were one year long or shorter, while the recessions for five states were at least 24 months long. Further, there were geographic patterns to the spread of the recession across states. I use these state-level estimates to introduce a new approach for calculating the total effects of recessions on state employment, one that accounts for lost employment growth as well the decrease in employment. States formed distinct geographic groupings according to these total effects, with states in the West and Southeast tending to have seen the greatest harm. Finally, many of the state-level differences in the effects of the Great Recession were related to differences in industry mix and the prevalence of sub-prime mortgages. The states with the longest and deepest recessions also tended to have been those with the highest shares of subprime mortgages.

本文针对大衰退（Great Recession）在不同州的发生时点、严重程度以及就业总影响的差异展开研究。研究发现，部分州早在官方认定的大衰退起始时间前便已陷入衰退，而另有十余州直至官方衰退起始六个月甚至更久后才进入衰退状态。各州的衰退结束时间同样呈现错落不一的特征。据此，11个州的衰退时长仅为一年及以内，而5个州的衰退时长则至少达到24个月。此外，大衰退在各州间的蔓延呈现出明确的地理分布规律。基于上述州级层面的测算结果，本文提出一种全新的测算方法，用于量化衰退对各州就业的总影响，该方法同时涵盖就业增长损失与就业规模缩减两部分效应。根据这类总影响效应，各州可形成特征鲜明的地理集群：美国西部与东南部各州往往受到的冲击最为严重。最后，大衰退在各州间产生的影响差异，大多与各州产业结构差异以及次级抵押贷款（sub-prime mortgages）的普及程度相关。衰退时长最长、影响最深的各州，往往同时也是次级抵押贷款占比最高的地区。