Severe information asymmetries are thought to make contracting particularly difficult within (and across) firms in developing countries. Standard principal-agent theory predicts that a new monitoring technology provided at zero cost should be widely adopted and unambiguously raise workers’ effort. I test this classical prediction using a field experiment with trucking companies in Liberia. The treatment offered to install GPS tracking devices on randomly selected trucks at no cost. Treatment-on-the-treated estimates reveal that the tracking devices increased monitored drivers’ average speeds by 58 percent, without leading to higher accident rates or maintenance costs. Despite this, managers declined to install the devices on 35 percent of the trucks selected for treatment. Using a model of intrinsic motivation, I show that it may be optimal not to monitor workers who are intrinsically motivated to work hard. While monitoring technologies increase agents’ extrinsic incentives to provide effort, they also do not allow worker to show that he or she does not need these incentives to work hard, which can crowd out effort. I provide three pieces of evidence in support of this explanation. First, Liberian trucking company managers choose to install tracking devices only on the trucks of drivers who perform less well at baseline. Second, the treatment effect on speed for monitored drivers is greater the lower the performance of the driver at baseline. Finally, I show that for drivers who performed well at baseline, the treatment has a negative effect on the relation between the manager and the driver, and on the driver’s propensity to follow the rules of the business. Overall, this paper demonstrates that, while new monitoring technologies can dramatically raise some workers’ productivity in settings where employment contracts are difficult to enforce, their use may lower the productivity of some workers – those who are intrinsically motivated to work hard.