Financial Intermediation without Exclusivity
研究金融中介在非排他性环境下（交易者可同时与多个中介交易）的均衡特征，发现竞争虽不能实现标准纳什均衡下的最优担保水平，但可通过预期均衡达到与排他性下相同的有效结果。
Futures exchanges and other financial intermediaries assume counterparty risks and, in return, demand guarantees that these counterparties will deliver on their promises. It is often argued that, to attract volume, financial intermediaries would settle for excessively low contractual guarantees. In Santos and Scheinkman (2001), we model financial intermediation in an environment where traders may choose to default, and we examine the characteristics of the equilibrium. In particular, we investigate whether competition implies excessively low standards. We show that, in fact, when society punishes default and intermediaries can impose collateral requirements effectively to limit the size of positions, competition leads to a (constrained) optimal amount of contractual guarantees. In Santos and Scheinkman (2001) we assume that exchanges cannot control the size of the positions taken by individuals, but we preclude investors from participating in more than one exchange. This ignores the effect that trading with one financial intermediary may have on the risks faced by other intermediaries. Financial intermediaries typically cannot control the amount of risk that counterparties will take with other intermediaries. In this paper we examine the effect of dropping the exclusivity assumption. We show that the constrained optimum can no longer be implemented as a standard Nash equilibrium with free entry. Nonetheless this allocation is the only one that can be sustained as an anticipatory equilibrium (Charles Wilson, 1977). To break a candidate anticipatory equilibrium it must be possible to add a contract that is profitable and that does not become unprofitable when the now unprofitable contracts from the original menu are withdrawn. If one views this requirement as reasonable, nonexclusivity reproduces the outcome that obtains when intermediaries enjoy exclusivity but cannot control the size of positions taken against them by traders. A superficially similar result holds in the insurance model of Richard Arnott and Joseph E. Stiglitz (1993). In that model, the price of insurance is simply the ratio of what the client pays in the favorable state versus what he receives in the adverse state. As a consequence, if insurers cannot control the size of policies, it makes no difference whether a client buys from one or many insurance companies: clients will buy as much as desired at the lowest price. As we will show below, in our model of financial intermediation, agents may want to combine several contracts, and this is precisely the reason why the original equilibrium does not survive as a Nash equilibrium once exclusivity is dropped.