Asserting control over the behavior of the judiciary represents one of the most important battlegrounds in the process of authoritarian consolidation. In this paper, we address the question of what accounts for resistance to authoritarian reforms aimed at bringing the behavior of judges in line with the regime's preferences. We present two sets of hypotheses: first, that resistance at the regional level is affected by the strength of local political opposition; and second, that resistance at the level of individual judges is facilitated by what we call 'embeddedness' in the local political environment, or personal and professional connections to local, rather than central, political actors. Measuring judicial compliance as decisions in favor of state-owned corporations or state entities, we test these hypotheses using a dataset containing the universe of decisions -- over 20 million individual cases -- made by regional commercial (arbitrazh) courts in the Russian Federation from 2008 to 2024. We focus on the case of a 2014 reform that abolished the higher commercial court, which we expect to have brought judges' behavior more in line with the state's preferences by overturning older precedents and centralizing legal decision-making. Due to the high volume of cases and arbitrary timing, we use a regression discontinuity in time framework to determine the causal effect of the reform and the sources of local heterogeneity.