Teacher absenteeism and shirking are common problems in developing countries. While monitoring teachers should ameliorate those problems, mobilizing parents to do so often leads to small or even negative effects on learning outcomes. This paper provides causal evidence that this might result from non-monotonic effects of monitoring teachers. Cross-randomizing nudges to teachers and parents in Ivory Coast - to motivate and monitor teachers directly, and to mobilize parents -, we find that, in schools where parents are nudged, numeracy and literacy test scores improve by an additional school quarter, and student dropouts decrease by over 50%. In contrast, in schools where both are nudged, there is no effect on either learning outcomes or dropouts - even though the latter also fall by nearly 50% where teachers are nudged alone. In those schools, teachers show up less frequently, allocate less time to career development, and target instruction to top students to a greater extent than in schools where only parents are nudged. Monitoring backfires precisely for teachers who were most motivated at baseline, consistent with monitoring intensity eventually crowding out intrinsic motivation. Our results have implications for the design of accountability programs above and beyond education.