Swapping Regulation for Litigation: The Effects of Workers' Compensation Opt-Out on the Injury Claims of Large Multi-State Firms

工人赔偿的“大妥协”，即工人放弃起诉雇主的权利，以换取有限的无过错职业伤害保险，是20世纪最重要的侵权改革之一。然而，由于工人赔偿在美国大多数州是强制性的和普遍的，很少有学者将其后果与侵权责任的后果进行比较。在方法上，问题是没有一个容易识别的“控制组”来比较受影响的工人和公司的行为。然而，有一个州--德克萨斯州--提供了一个独特的实验室来进行这种比较。根据德克萨斯州的法律，雇主可以选择不支付工人赔偿金，并丧失侵权豁免权。这种现象被称为“不认购”，自20世纪90年代以来，在大公司中一直在上升。有趣的是，尽管没有法律义务这样做，但大多数大的非认购者向他们的工人提供工伤计划，这些计划基本上复制了通过工人补偿提供的福利。这项研究是对德克萨斯州不认购的计划多部分调查的第一阶段，将研究不认购如何影响大公司职业伤害索赔的频率、成本和持续时间。任何这类研究都必须面对的关键方法论挑战是选择偏见。由于公司不是跨行业随机分配的，认购公司和非认购公司之间的公司水平差异可能反映了公司之间不可观察到的差异。为了最大限度地减少公司层面自我选择的潜在混淆影响，这项研究从在美国许多州运营同类设施的不同大型非订户群体中获得了高度细粒度的数据。为了利用各州之间的差异，这项研究将每个非订阅者在德克萨斯州设施中的员工与其非德克萨斯州设施中的员工进行了比较。与此同时，对于那些保留了取消订阅前一段时间数据的公司，调查员还将比较每家公司选择退出日期之前和之后的趋势。

The "great compromise" of workers' compensation, whereby workers relinquished the right to sue their employers in exchange for limited no-fault occupational injury insurance, was one of the most significant tort reforms of the 20th Century. Yet because workers' compensation coverage is compulsory and universal in most U.S. states, few scholars have compared its consequences to those of tort liability. Methodologically, the problem is that there is no readily-identifiable "control group" against which to compare the behavior of affected workers and firms. However, one state--Texas--offers a unique laboratory in which to undertake such comparisons. Under Texas law, employers may opt out of workers' compensation and forfeit tort immunity. This phenomenon, called "nonsubscription," has been on the rise since the 1990s among large firms. Interestingly, despite having no legal obligation to do so, most large nonsubscribers furnish their workers with occupational injury plans that substantially replicate the benefits provided through workers' compensation. This study, the first stage of a planned multi-part investigation into Texas nonsubscription, will examine how nonsubscription affects the frequency, cost, and duration of large firms' occupational injury claims. The key methodological challenge that any study of this type must confront is selection bias. Since firms are not randomly assigned across sectors, firm-level disparities between subscribing and nonsubscribing firms may reflect unobservable differences between firms. To minimize the potentially confounding effects of firm-level self-selection, this study has obtained highly granular data from a diverse group of large nonsubscribers that operate homogenous facilities in many U.S. states. To exploit variation across states, the study compares workers in the Texas facilities of each nonsubscriber to employees in its non-Texas facilities. Meanwhile, for those firms that have retained data from the period prior to nonsubscription, the investigator also will compare trends before and after each firm's date of opt-out.