The Supreme Court’s preemption jurisprudence over the last few decades has been unpredictable to say the least. Preemption defies traditional conservative–liberal alignment, as conservatives are torn between support of federalism and capitalist efficiency, and liberals are torn between support of strong national governance and multiplicity of legal remedies. As a result, justices often “flip” to take positions in tension with those that they take in federalism cases that do not involve questions of preemption. This tension between competing values combined with the complexity of preemption doctrine and the sheer number of values at stake in preemption cases produces great uncertainty. Court watchers and business interests alike have been unable to predict the path of the Court’s jurisprudence.

With its recent decision in Wyeth v. Levine, however, the Court has clarified its preemption analysis. Justice Thomas, in his concurrence, firmly rejects an entire line of the Court’s preemption decisions—“ purposes and objectives” obstacle preemption. This rejection is not a sudden development; Justice Thomas has disfavored obstacle preemption for some time. What is significant, though, is the opinion’s frankness and the clarity that it lends to the Court’s past (and future) conflict preemption analysis.

While urging the Court to abandon its obstacle preemption doctrine as overly broad and unsupported by congressional authorization, Justice Thomas simultaneously advocates for the expansion of what has traditionally been a very narrow category of preemption: impossibility preemption. Such an expansion is likely made necessary by the gap that would otherwise be left by his rejection of the obstacle preemption category. The natural question, then, is whether Justice Thomas’s expanded impossibility preemption might look very much like the narrow version of obstacle preemption that has been applied by the Court’s liberal bloc for some time. If this is the case, do we now have a firm five-Justice majority applying a unified doctrine to obstacle preemption cases? If so, what does that doctrine look like?

This Article sets out to answer these questions through an empirical analysis of the Court’s obstacle preemption decisions. Justice Thomas’s drift to disfavor obstacle preemption has been gradual. Although he did not formally renounce the doctrine until Wyeth, he has presumably been applying a similar analysis for some time. An examination of each of the Court’s obstacle preemption cases over the last fifteen years confirms that presumption. The analysis shows that Justice Thomas’s decisions very closely parallel those of the Court’s liberal bloc over the same time period. His broad impossibility analysis is thus functionally coterminous with a narrow version of obstacle preemption. Justice Thomas and the Court’s liberals form a distinct and somewhat reliable anti-obstacle preemption bloc. The analysis also shows, however, that because Justice Thomas and the Court’s liberals arrive at their positions through quite different analyses, their bloc is subject to fracture in certain particularly contentious cases.

Part II presents a brief overview of the Court’s current preemption doctrine. Part III details the breakdown of traditional left–right voting blocs in preemption cases and the consequent failure of political voting models to explain the Court’s preemption jurisprudence. Part IV uses the recent Wyeth v. Levine decision as a lens through which to view the Court’s peculiar voting patterns in obstacle preemption cases. Part V presents an empirical analysis of the Court’s obstacle preemption cases over the last fifteen years. The data show an emerging five-Justice voting bloc opposed to obstacle preemption resulting from the surprising alignment of Justice Thomas with the Court’s liberal wing. Finally, Part VI identifies the factors that the loosely aligned voting bloc finds most relevant to its preemption decisions, exposes weak points in the bloc where fracture is likely, and discusses how recent appointments will affect voting alignments.