Political scientists and legal scholars debate the value of judicial elections, including the degree to which elections effectively hold incumbent judges accountable. In this paper, we provide a causally identified estimate of the incumbency advantage in judicial elections. We assemble an original dataset of over 5300 partisan, single-member trial court elections from six U.S. states. Employing a regression discontinuity design, we demonstrate that incumbents enjoy electoral advantages of more than twenty percentage points due solely to being an incumbent. In contrast to research from other electoral settings, we find that these advantages are due largely to a scare-off effect, where even a narrow victory dramatically decreases the probability that an incumbent party will be challenged in the next election. Our findings highlight the sizable electoral returns to holding judicial office, reveal how the nature of the incumbency advantage varies across electoral settings, and provide compelling evidence of the challenges to holding trial court judges accountable through elections.