The primary functional justifications for eminent domain involve bargaining problems, including the holdout problem, the bilateral monopoly problem and other transaction costs, as well as the existence of externalities. The holdout problem is particularly noteworthy, and this chapter analyzes three types of holdouts, depending on whether the failure in bargaining is the result of strategic behavior among owners, the presence of a large number of owners or a single owner who is unwilling to sell because of a highly idiosyncratic valuation. Although eminent domain solves any potential bargaining problems by transferring land directly from existing owners to the government, eminent domain has limitations as well. The primary limitations are the difficulty of valuing parcels, the potential for secondary rent seeking and the existence of administrative costs. Valuing parcels is especially problematic because, in the absence of perfect information, the government may underestimate the valuations of existing owners or overestimate the valuations of future owners. In either case, eminent domain may increase the likelihood of an undesirable transfer, i.e., a transfer in which the existing owners value the land more than the future owners, even if public officials are acting to maximize social welfare rather than advance their own, or other private, interests. Because of these limitations, scholars have proposed various alternatives for acquiring land, three of which I discuss below: (i) secret purchases; (ii) land assembly districts; and (iii) auction mechanisms. Comparing eminent domain with each of these alternatives is necessary to determine whether, or under what circumstances, it is desirable for the government to invoke eminent domain, a determination that ultimately depends on a number of important, yet relatively unexplored, empirical questions
