The effects of accountability to constituents do not have to be all bad, however. Constituents can keep negotiators from making extreme or outrageous commitments that might get them in trouble later. For example, Kirby and Davis (1998) had constituents monitor the investment decisions of managers in a simulated production game. The results of the experiment indicated that monitored managers were less likely than unmonitored ones to escalate their commitment to unproductive courses of action and less likely to pursue risky investment strategies. Thus, accountability can deter individuals from pursuing risky decisions risky decisions that may have long-term destructive consequences. And Fassina (2004) points out that constituents can clearly develop at least two different kinds of contracts with their agents. The first is a behavior-contingent contract, in which the agent is primarily paid based on how he or she behaves in the role, versus an outcome-contingent contract, in which the agent is primarily paid based on the type of results he or she achieves. “Combined” contracts that specify both behaviors and outcomes are also possible. Fassina shows how aspects of the specific negotiation—expertise, emotional strain, the zoon of possible agreement and other issues—can help determine which from of contract is preferable and how those different contracts might lead to different negotiating behaviors displayed by the agent. Some clear ways for holding a negotiator accountable are offered in Box 11.4.