Alan S. Blinder–John Morgan collaboration

The Alan S. Blinder–John Morgan collaboration examined whether small groups could make intelligent decisions without sacrificing speed. In experiments designed to resemble the Federal Reserve’s challenges, groups reached more accurate answers than individuals, including the strongest individual in the group. The results support a distinction central to the book’s account of collective judgment: assembling people is not enough; the decision process must give their separate information a way to contribute to a shared answer.

Why Blinder and Morgan Tested Groups

In 2000, Princeton economists Alan S. Blinder and John Morgan designed a study to test two familiar doubts about committee decisions: whether groups could decide intelligently and whether they could do so quickly. Blinder had served as vice chairman of the Federal Reserve Board in the mid-1990s, and that experience had left him skeptical of decision making by committee. The Federal Open Market Committee, which sets interest-rate changes, consists of twelve members: the seven members of the Federal Reserve Board and five presidents of regional Federal Reserve banks. The study used two deliberately simplified experiments to approximate the Fed’s task of detecting changes in economic conditions and responding appropriately.

The design made speed part of the test rather than treating it as an afterthought. In the first experiment, students first made decisions alone, then discussed and decided as a group, returned to individual decisions, and finally worked together again. This sequence was intended to control for learning: any apparent group advantage could not simply be attributed to participants becoming more familiar with the task over time.

Detecting a Change in the Evidence

The first experiment placed equal numbers of red and blue balls in urns. Participants knew that, sometime after the first ten draws, the proportions would shift, leaving either 70 percent red and 30 percent blue or the reverse. Their task was to identify as soon as possible which color had become more prevalent. The setup modeled the problem of recognizing that economic conditions had changed and deciding whether a shift in monetary policy might be needed. To make prompt recognition matter, participants were penalized for every draw they made after the changeover.

The groups proved both faster and more accurate than individuals. They identified the direction of the shift correctly 89 percent of the time, compared with 84 percent for individuals, and they outperformed even the best individual participant. The result is more striking because the experiment asked the group to act under pressure: accuracy was not gained by simply taking more time to deliberate. In this task, discussion helped participants identify a change in the evidence quickly enough to avoid the cost of delay.

Setting Interest Rates as a Group

The second experiment asked participants to play the role of central bankers, setting interest rates in response to changes in inflation and unemployment. They had to detect whether the economy was slowing or gathering speed and move rates in the appropriate direction. Again, the groups made better decisions than individuals, who more often moved rates the wrong way. Yet the groups were just as quick as the individuals, answering the objection that committee deliberation necessarily makes a decision process inefficient.

The comparison with the strongest member sharpened the finding. There was no correlation between the performance of the smartest person in a group and the performance of that group. The collective result was not merely the best participant’s judgment passed along to everyone else: the group made better decisions than its smartest members. A Bank of England study modeled on Blinder and Morgan’s experiment reached the same conclusions, finding that groups could decide intelligently and quickly and could outperform their best members.

What Makes the Group Decision Count

The experiments’ significance lies not only in showing that groups can perform well, but in the conditions under which their performance becomes useful. In another classroom experiment, Angela Hung and Charles Plott gave participants private evidence in the form of colored marbles drawn from one of two urns. One urn contained twice as many light marbles as dark ones; the other contained twice as many dark marbles as light ones. Participants drew in sequence, kept the color of their own marble private, and announced only which urn they thought was being used. When three people ahead guessed Urn B, for example, a participant who had drawn a light marble faced a conflict between private evidence and the visible pattern of earlier answers.

Information cascades began in 78 percent of the trials. Hung and Plott then changed the incentive: instead of earning money for an individually correct answer, participants were paid according to whether the group’s majority answer was correct. The shift made the group’s accuracy, rather than each participant’s individual score, the object of the task. Participants paid more attention to their own imperfect evidence and less to the guesses of those who had already answered. Individual guesses could therefore be wrong more often while the group as a whole became more accurate than the cascading groups.

The comparison clarifies why the order of decisions matters. Sequential choices let early answers influence later ones, so private information can disappear behind an apparent consensus. Hung and Plott’s changed reward reduced the importance of earlier choices and encouraged independent judgments; it did not make each person’s evidence perfect. Instead, it gave the group a better chance to combine imperfect information. The book draws a broader lesson from that result: organizations can often improve decisions by eliciting judgments simultaneously, or nearly so, rather than making members speak one after another.

Committees Need a Way to Decide

Blinder and Morgan’s findings challenge the assumption that a small group is inherently slower or less capable than its individual members. They also set a condition on the value of teams: a group that offers only advice, without any method for aggregating its members’ opinions, cannot realize the advantage the experiments demonstrate. If an organization includes small groups in decision making, the group must be allowed to make decisions rather than merely report separate views for someone else to consider.

The Columbia case supplies the book’s pointed contrast. The Mission Management Team did not vote on anything: members reported on different aspects of the mission, but their opinions were never aggregated. The book treats that absence of a collective decision as a mistake regardless of whether the shuttle had returned safely. Blinder and Morgan’s experiments thus support neither blind faith in committees nor a simple preference for individual authority. They show that collective intelligence can arise quickly when a group is given a task and a means of turning its members’ judgments into an actual decision.