In this paper we analyze the effect of the design of yardstick competition on consumer prices, by means of a theoretical analysis as well as an economic experiment. We compare four different designs: the uniform yardstick, the unweighted uniform yardstick, the discriminatory yardstick, and the best-practice yardstick. The effect of a specific design on prices depends on two separate mechanisms, one which affects the incentive power to increase productive efficiency and another which affects the risk of collusion. We show theoretically that for the best-practice yardstick, which is widely applied in several industries in a number of countries, these two mechanisms point in the same direction (high prices), which is confirmed by the experiment. Both the theoretical analysis as well as the economic experiment show that the discriminatory yardstick results in lower prices than the unweighted uniform yardstick. The theory, however, does not give a clear answer on the relative performance of the discriminatory versus the weighted uniform yardstick. In the experimental analysis, we find that the advantage of the discriminatory yardstick in terms of giving incentives to improve productive efficiency exceeds the disadvantage of a relatively higher risk of collusion. This conclusion appears to be robust for different degrees of heterogeneity of the industry. Hence the discriminatory yardstick yields the lowest prices for consumers. (C) 2017 Elsevier B.V. All rights reserved.