Abstract: In the past two years, the Supreme Court has invalidated two major executive-branch initiatives—the termination of the Deferred Action for Childhood Arrivals (DACA) policy and the addition of a citizenship question to the census—as arbitrary and capricious. Many have cast Chief Justice Roberts’s decisive votes and opinions in these cases as efforts to protect the Court’s public standing by skirting political controversy. Taken on their own terms, however, the opinions seem less about keeping the Court out of the political thicket and more about pushing the Trump Administration into it. And that use of arbitrariness review as a judicial backstop for political accountability is an important jurisprudential development in its own right. For decades, the Court has understood arbitrariness review mainly as a check against bureaucratic blunders, lawlessness, and political interference with agency expertise. But in the DACA and census cases, a narrow majority refashioned this form of review as a tool for forcing an administration to pay the appropriate political price for its discretionary choices. Through close and context-laden readings of these back-to-back opinions, I aim to surface the “accountability-forcing” form of arbitrariness review that they employ and to draw out its significance. Between the two cases, the Roberts-led majority identified three kinds of agency explanations that should be rejected or disfavored on political-accountability grounds: post hoc explanations, buck-passing explanations, and pretextual explanations. Standing alone, these new rules (and new justifications for old ones) have wide-ranging consequences. But if the shift toward an accountability-centric vision of arbitrariness review continues, it could also lead to renovations of several other administrative-law doctrines—including narrowing the carve-outs from judicial review, undermining the remedy of “remand without vacatur,” and empowering courts to discount agencies’ fallback justifications for their choices. After laying out the accountability-forcing turn in the Court’s recent cases and sketching its possible ramifications, I consider several grounds for doubt about its propriety and efficacy. Some of these objections, I conclude, have real force. Still, none debunks the core insight that I take to underlie Roberts’s approach: The reasoned explanation requirement can sometimes be deployed so as to promote not only rational administration, but democracy as well.