In incomplete information environments with transferable utility, efficient outcomes are generally implementable unless interim or ex post participation constraints are imposed on the problem. In this paper we show that linking a sufficiently large number of independent but possibly unrelated social decisions, a slightly perturbed Groves mechanism can implement an efficient outcome with probability arbitrarily close to one, while respecting all participation, incentive and balanced budget constraints.
This paper studies the role of bundling in the efficient provision of excludable public goods. We show that bundling in the provision of unrelated public goods can enhance social welfare. With a large number of goods and agents, first best can be approximated with pure bundling. For a parametric class of problems with binary valuations, we characterize the optimal mechanism, and show that bundling alleviates the free riding problem in large economies and decreases the extent of use exclusions. Both results are related to the idea that bundling makes it possible to reduce the incidence of exclusions because the variance in the relevant valuations decreases.
This paper studies the optimal provision mechanism for multiple excludable public goods when agents’ valuations are private information. For a parametric class of problems with binary valuations, we characterize the optimal mechanism, and show that it involves bundling. Bundling alleviates the free riding problem in large economies in two ways: first, it can increase the asymptotic provision probability of socially efficient public goods from zero to one; second, it decreases the extent of use exclusions.
Keywords: Public Goods Provision; Bundling; Exclusion
Commodity bundling is studied in an environment where the dispersion of valuations unambiguously decreases when two or more goods are sold as a bundle only. Bundling is more likely to dominate separately selling the goods if marginal costs are low relative to the average valuation, or if the distribution of valuations is very peaked around the mean.