Format of Original
Utilizing experimental data on choices over real monetary rewards made by university students, we provide evidence that two measures of liquidity, income and employment status, significantly explain differences in patterns of discounting. We find an average fixed cost component of discounting in the range of $5 for unemployed students and near $0 for employed students. An increase in annual disposable income of $1000 decreases the fixed cost component of discounting by approximately $0.20 to $0.25. These findings can help resolve the puzzle that some studies in the literature find evidence of present-bias and magnitude effects and some do not.