There has been an extensive research literature on auctions, but recent developments in technology have resulted in new interest in auction mechanisms as a practical way of allocating resources. This paper presents a new double-auction mechanism to handle resource allocation for public goods when complementarity exists. The mechanism is placed in the context of an organization's internal knowledge investment. Knowledge goods have two distinct characteristics. First, knowledge within an organization can be considered a public good, so it is subject to the free-rider problem. Second, knowledge is interrelated and interdependent; that is, there is complementarity among knowledge components. The value of knowledge often derives from a bundle of knowledge components, rather than from its individual pieces. These two characteristics present a serious challenge to allocating organizational resources for knowledge goods. We introduce an internal market in which knowledge providers offer knowledge projects and knowledge consumers place bids to acquire them. The mechanism is a Groves-Clarke-type double auction that allows bundled knowledge goods to be traded so as to recognize complementarities between knowledge projects. The market mechanism we propose is incentive compatible; i.e., it induces people to reveal their true valuation. In addition, it allows trades of knowledge bundles to determine which knowledge components are most valuable from the organization's viewpoint. Under mild assumptions, the mechanism is a computationally tractable solution to operating a market of bundled public goods. We further show how “imputed prices” can be calculated for subsets of knowledge components and prove that a market mechanism that does not allow bundle orders or does not address the free-rider problem yields a systematic underinvestment in knowledge.
Stored only when both papers are in the corpus — 103 citations exist upstream.