This laboratory experiment studies two-stage contests between political parties. In the first stage, parties run their primaries, and in the second stage, the winners of the primaries compete in the general election. The resource expenditures in the first stage by the winning candidates are partially or fully carried over to the second stage. Experimental results support all major theoretical predictions: the first-stage expenditures and the total expenditures increase, while the second-stage expenditures decrease in the carryover rate. Consistent with the theory, the total expenditures increase in the number of candidates and the number of parties. Contrary to the theory, however, expenditures in both stages of the competition exceed theoretical predictions. Disclosing information about the opponent’s expenditures in the first stage increases the second-stage expenditures and decreases the first-stage expenditures.