The authors extend research and theory on self prediction into the realm of personal financial behavior. Four studies examined people's ability to predict their future personal spending and the findings supported the two main hypotheses. First, participants tended to underestimate their future spending. They predicted spending substantially less money in the coming week than they actually spent or than they remembered spending in the previous week. Second, the prediction bias stemmed from people's savings goals-defined as the general desire to save money or minimize future spending-at the time of prediction. Participants who reported (Studies 2 and 3) or were induced to experience (Study 4) a stronger savings goal predicted they would spend less money. However, savings goals were not related to actual spending and thus contributed to the bias in prediction.