Monetary incentives can backfire while nonstandard interventions, such as framing, can be effective in influencing behavior. I review the empirical evidence on these two sets of anomalies. Paying for inherently interesting tasks, paying for prosocial behavior, paying too much, paying too little, and providing too many options can all be counterproductive. At the same time, proper design of the decision-making environment can be a potent way to induce certain behaviors. After presenting the empirical evidence, I discuss the relative role of beliefs, preferences, and technology in the anomalous impacts of incentives. I argue that inference, signaling, loss aversion, dynamic inconsistency, and choking are the primary factors that explain the data.