Sequential advocacy

The collection of information necessary fordecision-making is often delegated to agents (e.g. bureaucrats,advisors, lawyers). If both the pros and cons of a decision haveto be examined, it is better to use competing agents instead of asingle agent. The reason is that two conflicting pieces ofinformation cancel each other out. Using two agents, eachsearching for one cause yields full information collection atminimum costs. This provides a rationale for advocacy in politicaland judicial systems. In this paper, we provide a rationale forthe sequential nature of information collection in advocacysystems. If two agents search simultaneously, the incentive tocontinue searching is affected by the information found by theother agent. This forces the principal to leave rents to theagents. If agents search sequentially, the reward can be madeconditional on the information found in earlier stages. Thisreduces the cost of information collection. However, sequential advocacyimplies either a more sluggish decision-making processor a less-informed decision.