The undeciders

It’s the other Parkinson’s: the progressive degeneration of a committee’s ability to make decisions as the committee adds more members.

English historian Cyril Northcote Parkinson observed in the 1950s that decision making is severely impaired in committees of more than 20 people. Now physicists have shown that the size of a country’s executive cabinet appears to be linked to that country’s overall efficiency, and they have found a possible mathematical explanation.

Stefan Thurner, a physicist at the Medical University of Vienna, and his collaborators looked at the overall efficiency of virtually every government on the globe, as measured by United Nations and World Bank indicators taking into account factors such as literacy, life expectancy and wealth.

The researchers then looked at each country’s executive cabinet. “Cabinets are a good representation of countries,” Thurner says. Common sense would suggest that smaller cabinets would find it easier to reach a consensus. But to get the rest of the country behind a decision, cabinets also have to be large enough to represent of a wide range of constituencies, Thurner says. “Behind every minister there is a set of lobbyists, interest groups and a large bureaucracy.”

On average, the team found, a country’s development was tied to the size of its executive cabinet. For example, Iceland, which the United Nations ranks as the world’s most developed country, has a cabinet of just 12 members; the United States, which ranks 12th, has 17 cabinet members; Myanmar and the Ivory Coast, with 35-strong cabinets, rank 132nd and 166th^.

The researchers also tried to figure out exactly how a committee’s size affects its efficiency and to explain Parkinson’s 20-person rule.

The team simulated committees as networks in which each member was a node. Before a vote, each member’s opinion could be influenced by those of its immediate neighbors in the network; adjacent nodes could represent, for example, ministers belonging to the same political party. The simulation found that committees of 10 members or less could almost always reach a consensus (with one mysterious exception for the number 8). For larger committees, the chances of getting to a consensus were lower, and the chances decreased even more rapidly for committees of 20 or more. The results show that Parkinson’s law is not an accident, but “a robust consequence of the opinion-formation model,” Thurner says.

“It’s interesting that they find a correlation,” says Yaneer Bar-Yam of the New England Complex Systems Institute in Cambridge, Mass. However, Bar-Yam points out that the correlation is only true on average. In fact, the data show some important exceptions. For example, Australia, Canada and New Zealand have large cabinets but high efficiency scores. A committee’s effectiveness, Bar-Yam says, strongly depends on how the committee is organized. “One of the great examples today is Wikipedia,” he says. The online encyclopedia manages to function despite being written and edited by thousands of volunteers because of the way it’s structured, he says.