Intermediaries as information aggregators : an application to US treasury auctions

According to most theories of financial intermediation, intermediaries diversify risk, transform maturity or liquidity, and screen or monitor borrowers. In U.S. Treasury auctions, none of these rationales apply. Intermediaries submit their customer bids without transforming liquidity or maturity, and they do not screen or monitor borrowers or diversify fiscal policy risk. Yet most end investors place their Treasury auction bids through an intermediary rather than submit them directly. Motivated by this evidence, we explore a new information aggregation model of intermediation. Intermediaries observe each client's order flow, aggregate that information across clients, and use it to advise their clients as a group. In contrast to underwriting theories in which intermediaries, by acting as gatekeepers, extract rents but reduce revenue variance, information aggregators increase expected auction revenue but also make the revenue more sensitive to changes in asset value. We use the model to examine current policy questions, such as the optimal number of intermediaries, the effect of non-intermediated bids, and minimum bidding requirements.