Specific investments, holdup, and the outside option principle: an experimental study

According to the outside option principle the holdup problem can be lifted by structuring the post-investment bargaining stage in such a way that the non-investing party has a binding outside option. The investor then becomes residual claimant and has the appropriate incentives to invest. Theory predicts that investments increase when noninvestors' outside options increase from a non-binding low level to a binding high level. Investments in the latter case are then equal to the socially efficient level. Further, it is also predicted that this mechanism does work when breakdown is optional (outside option game) and does not work when there is only forced breakdown of the bargaining (threat point game). This paper reports about an experiment designed to test these predictions. We find that investment levels do not increase with the value of the non-investor’s outside option. When the outside option is binding investment levels fall short of the socially efficient level. But, in outside option games investment levels tend to be higher than in threat point games. Actual investment levels can be explained by actual bargaining results which also deviate from the predictions based on subgame perfectness. When theory predicts investors to receive half of the surplus created by the investment, the agreed amount exceeds the predicted amount. When theory predicts investors to be residual claimant, the agreed amount falls short of the predicted amount. The main difference between the outside option game and the threat point game appears not to be the operation of the outside option principle, but rather that agreement is reached sooner. Overall, our experiment suggests that the effectiveness of contractual solutions to holdup that rely on the outside option principle is rather limited in practice. JEL Codes: J41, J31, C91 * An earlier version of this paper was presented at a seminar at the University of Amsterdam in september 1999, and at the European regional meeting of the ESA in Grenoble, October 1999. Helpful comments by Laurent Denant-Boemont, Edwin Leuven, Theo Offerman, and Arno Riedl are gratefully acknowledged. The usual disclaimer applies.