CHEVRON AS A DOCTRINE OF HARD CASES

According to the conventional wisdom, the Chevron doctrine rests on a presumption about congressional intent-a presumption that when a statute is ambiguous, Congress intended the gap to be filled by the agency charged with administering the statute. But the presumption is a mere fiction; when Congress enacts a statute, it generally has no view on who should resolve the ambiguities that later arise. This Article proposes a new theory of Chevron, one that rests on a simple reality: no matter how determinate the law may seem, there will inevitably be hard cases-cases in which the law runs out before providing a solution. As legal positivism teaches, hard cases cannot be decided by merely applying existing law. When the law runs out, a case can be decided only by making new law to fill the gap. There are thus two distinct stages in deciding every hard case: applying the law and making it. This Article argues that these two stages correspond to Chevron's two steps. Step One is the ordinary, law-applying stage of any case of statutory interpretation. Step Two is the law-making stage, when a court is faced with a gap to fill. The presence of an agency construction, however, means that the court itself need not make law to fill that gap; instead, it may defer to the law-making of the agency-which, unlike the court, is accountable to the political branches. Viewed this way, deference emerges as an act of judicial self-restraint, grounded in the recognition that the law carries greater legitimacy when made by politically accountable agencies than by unelected judges. This positivist account of Chevron elucidates the doctrine's familiar two-step inquiry, shedding light on longstanding questions about the doctrine's application. It also answers recurring objections to judicial deference more generally, including the claim that such deference conflicts with the Constitution. Finally, understanding Chevron as a doctrine of hard cases has important implications for the scope of Chevron's domain.