Williamson's The Economic Institutions of Capitalism

* There can be no doubt-this book is a major contribution. Perhaps the volume is best characterized as a systematic theoretical examination of those portions of society's economic activities that the mechanism does not automatically run to perfection. But here such activities are not simply listed as market failures and forgotten. Rather, they are treated, appropriately, as normal and widespread economic phenomena whose anatomy urgently requires examination and understanding. Vertical integration is an archtypical example. If the unaided mechanism did everything perfectly, firms would rarely be able to enhance efficiency by integrating vertically. It would usually be optimal for intermediate goods to be bought from independent suppliers on the free market. I have more than once enjoyed teasing businesspersons bent on defending the vertical integration of their firms by suggesting that they are Soviet central planners at heart and that, when push comes to shove, they have little use for the mechanism. In practice, of course, the truth generally lies somewhat in between. For any given firm there is presumably an optimal degree of vertical integration determined by the requirements of economy in production and economy in the governance process, or as Williamson puts it, of economizing in transactions costs. The point is that a variety of institutions and behavior patterns really make sense primarily in terms of their relative efficacy in dealing with transactions costs that renders each, in turn, best in different circumstances: decentralized and hierarchical organization of firms, the vast variety of contractual forms, franchising, tie ins, and many others. These are the sorts of institutions to which Williamson's analysis is devoted. It should be said at once that I can find no major point of disagreement with his analysis. On the contrary, I enthusiastically accept the substance of all his central conclusions. This extends also to Williamson's characterization of the contestability literature and its relation to his own analysis which, I believe, he has done completely correctly, as will be spelled out presently. I have only two substantive reservations about the book. The first is a slight tendency by the author to overstatement of claims for the new analysis (a manifestation of enthusiasm which is no doubt commendable in a pioneer, and, in any event, I know I am hardly in a