This paper examines U.S. Office of Management and Budget (OMB) Circular A-94, which requires that costs and benefits of government programs be discounted at the U.S. Treasury borrowing rate on bonds of maturity equal to the period of analysis. It is shown that, in a 2-period problem, discounting at the Treasury rate always leads to the correct investment decision. That is, the winning project may always be transformed, via market transactions, into one that dominates the losing project. Some anomalies occur when evaluating projects of longer duration, but these anomalies are probably not important empirically. Finally, the dominating project may be computed using linear programming.