Donald Trump repeatedly vowed to reduce regulation during the 2016 presidential campaign. Indeed, one of his key advisors promised to "deconstruct" the administrative state. Since taking office, President Trump has attempted to make good on his promises, spurring federal agencies to brush aside countless regulations that previous administrations had promulgated based on scientific, technological, or economic evidence. Those efforts, which have been dubbed a "war on science," implicate a long-contested question in administrative law: to what extent should a change in presidential administrations excuse agencies from an obligation to justify changes in policy with expert, reasoned analysis of relevant data? Perhaps surprisingly, the Trump administration's efforts align with views that have dominated administrative law scholarship in recent decades. By the time Trump took office, many leading administrative law scholars had already championed enhanced presidential control over agency decisions, dismissed expert analysis as an anachronistic relic of the New Deal, and suggested that the considered judgments of previous administrations should be amenable to quick and easy change. This article takes a contrary view and asserts a renewed role for expert, reasoned analysis in the face of politically motivated administrative change. Unlike earlier work, this article identifies change as a fundamental and essential aspect of much expert decision-making, and it explains that regulatory statutes often call for an exercise of expert judgment capable of incorporating frequently changing bodies of scientific, technological, or economic knowledge. This positive procedural account of agency decision-making shows that the reasoned analysis contributed by agency expertise is far from superfluous, but contributes legitimacy and transparency to administrative government. By identifying the value of expertise within the context of politically directed policy changes, this article addresses an under-theorized aspect of judicial review of agency decisions and reinforces the need for agencies to support changes in policy with reasoned, expert analysis.