Groupthink or Deadlock: When Do Leaders Learn from Their Advisors?

1. Introduction

CHAPTER 1 Introduction It may be lonely at the top, but hardly ever so lonely that important decisions in government and business are made by only one person. Even presidents and prime ministers must rely on others. In fact, the most powerful leaders generally confront such a range of problems that they require assistance and advice more frequently than less prominent figures. Small wonder, then, that the most important and time-consuming issue confronting newly elected presidents before they take office is not to determine their own position on various policy issues, but simply to decide how they will organize their staff. In business as well as in government, this is a crucial problem . Schools of management debate ad nauseam the most effective and efficient way to fashion work groups of various sizes and functions. Does the Japanese ringi system of collective decision making, for example, produce better decisions than Western management styles even though it may seem less efficient?1 Are cohesive groups more productive, or is some disagreement healthy?2 How important is strong leadership for group efficiency?3 Such questions are the staples of research on group decision making. And while their answers are debated, almost everyone agrees that groups deserve attention in their own right. Yet no matter how indispensable a well-organized staff may be, even carefully assembled groups of advisors can sometimes 1 create more problems than they solve. Many studies of presidential decision making argue that social pressures within high-level policy groups can lead presidents to make worse decisions than had they acted alone. The best-known work of this genre, Irving Janis’s study of groupthink, holds that this danger is especially prevalent in highly cohesive groups.4 Janis suggests that taking steps to prevent “excessive” conformity and to promote “devil's advocacy” will make the group a more benign setting. Janis was not the first to recognize the danger of groupthink. It was probably common wisdom even when Aristotle wrote that “evils draw men together.”5 The danger is not simply that the presence of like-minded others can reinforce one’s own sense of rectitude and virtuousness far beyond prudence. The mere presence of others is energizing in a way that can prompt rash action or even mob behavior.6 Groups also provide a measure of anonymity; they make it easier to entertain the belief that, when plans go awry, someone else can be blamed.7 This aspect of groups not only encourages risky behavior, but it creates problems of accountability that undermine democratic procedures. It is bad enough that leaders might be led by advisors to contemplate unwise adventures , but even worse when leaders purposefully distance themselves from the details of an operation to preserve “deniability,” thus subverting their accountability before the public.8 One is tempted to assert that leaders (and their constituents) would be better off by themselves than in a room full of advisors.9 Whatever its merits, no policymaker could really afford to adopt this maxim. Even the most knowledgeable leaders cannot be experts on everything. Neither side of the debate over the usefulness of advisors really takes issue with an underlying premise: that more information is always better, so long as there is sufficient time to consider it. Information, one might say, is the lifeblood of decision making. The assumption that it is intrinsically valuable is a truism among almost all of those who study group decision making. MATCHING ADVISORS TO LEADERS This book rejects the assumption that more information is always helpful—even when the information itself is important and even when there is sufficient time to consider it. For people as for com2 Introduction Introduction 3 puters, information is not helpful if it arrives in too great a quantity to process. For computers, the problem is simply one of processing speed. All that is necessary, for computers, is enough time. But for people, the problem is processing speed and tolerance for ambiguity, diversity of opinion, complexity, and many other features of the information itself. Computers are indifferent to the information they process, but people obviously are not. To store knowledge efficiently, the brain simplifies and distorts learned information. One of the most basic assumptions of cognitive psychology is that people attempt to preserve consistency among their beliefs, fitting beliefs to simpler (and thus more efficient) mental frameworks.10 When we are unable to preserve consistency, however, we do not...