Delay after disagreement in committee decision making may foster information aggregation but is costly ex post. When there is an upper bound on delay that can be credibly imposed, repeated delays can improve the ex ante welfare of committee members. An ex ante optimal dynamic delay mechanism does not impose the maximum credible delay after each disagreement. Instead, it induces in equilibrium start-and-stop cycles where players alternate between making the maximum concession to avoid disagreement and making no concession at all. The start-and-stop feature is robust to modeling delay cost by discounting instead of money-burning, and the optimal mechanism is shown to be “redesign-proof” when there is also an upper bound on the number of rounds of delay that can be credibly imposed.

在委员会决策中出现分歧后的延迟可能会促进信息聚合，但事后代价高昂。当存在可信的延迟上限时，反复延迟可以改善委员会成员的事前福利。事前最优动态延迟机制不会在每次分歧后施加最大可信延迟。相反，它引发了平衡的开始和停止循环，玩家在做出最大让步以避免分歧和根本不做出让步之间交替。启动和停止功能通过折扣而不是烧钱对延迟成本建模具有鲁棒性，并且当延迟轮数也有上限时，最佳机制被证明是“重新设计证明”的被可信地强加。