Democracy can be better served by randomly selecting representatives, argue Italian researchers.
Dr Alessandro Pluchino of the Università di Cantania and colleagues report their findings on the pre-press website arXiv.org.
"We think that the introduction of random selection systems, rediscovering the wisdom of ancient democracies, would be broadly beneficial for modern institutions," write the researchers.
Pluchino and colleagues developed a computer simulation, in which they studied the behaviour of politicians when randomly selected independents were introduced to a model parliament.
Their model relied on four categories of people in the parliament. These were: 'intelligent' people (actions serve both personal and social interests), 'helpless or naive' (loss for self, but gain for others), 'bandits' (benefit themselves, but not others), and 'stupid' (actions produce a loss for everyone).
The model, involving a parliament made up of two parties, had 500 individuals who could each propose and vote for or against acts.
Pluchino and colleagues found that in all cases studied, adding random legislators improved the performance of the parliament. Specifically, there were more acts passed with social benefit.
"Most people think that democracy means elections," they write. "However ... in the first significant democratic experience, namely the Athenian democracy, elections worked side by side with random selection (sortition) and direct participation."
Pluchino and colleagues say the drawbacks of a system dominated by political parties have been well documented. These include the tendency for politicians to follow a 'party line' and the tendency of groups to defend their interests.
In recent decades, they write, there has been a rise in interest in choosing representatives by random selection.
Arguments in favour of this are that it can reduce corruption, prevent the dominance of a small group of politically active people, and ensure people of different incomes, races, religions and sex, are more fairly represented in parliament.
In 2010, Pluchino and colleagues won an IgNobel prize for showing mathematically that organisations could avoid the spread of incompetence - or the Peter Principle - and become more efficient if they promoted people at random.
Limits to random selection?
Australian politician, Andrew Leigh describes the new research on random selection as an "interesting contribution".
But he sees random selection as being useful in limited situations - such as in deliberative polling, prior to a referendum.
"In a country that struggles to get people to perform a couple of days jury service, the notion that randomly-sampled people would be happy to give up their day jobs and become legislators seems far-fetched," says Leigh.
Leigh, a member of the Australian Labor Party and, until recently a professor of economics with an interest in public policy at the Australian National University, says Pluchino and colleagues are using an approach related to game theory and political science. Game theory is a branch of maths that tries to model behaviour of individuals making choices.
But Leigh challenges some of the assumptions in the parliamentary model.
"I think it's a mistake to regard many politicians as having taken up the profession because they're self-interested," he says.
"[Pluchino and colleagues] speak about bandits in their model - bandits are there to enrich themselves. I don't see any bandits in Australian politics," says Leigh.
He says it is more appropriate to think of politicians as having different views about the world rather than some being good and some being venal.
Wisdom of the crowd
Professor Lyn Carson of the Centre for Citizenship and Public Policy at the University of Western Sydney says random selection can improve deliberation, as well as representativeness, in democracy.
But she says most political science models of sortition tend to assume the whole parliament is randomly selected.
Carson says Pluchino and colleagues' model is different because it is proposing the introduction of just a few randomly -selected independents.
"I like that idea," says Carson, who was involved with the 2009 Australian Citizens' Parliament.
But like Leigh, she is not persuaded by Pluchino and colleagues' categories of individuals in parliament.
She says these categories may exist in adversarial parliamentary systems, but not when there is a proper deliberation.
"There is a very dramatic shift at a point in the deliberative environment when people shift from self interest to public interest," says Carson.
She says random selection is used to ensure fairness in a diverse range of areas from the allocation of the US Green Card to social housing in Ireland.
"We do it because it's fair. Everyone has got an equal chance," says Carson. "It doesn't make sense to exclude parliaments from that."
She says good democratic deliberation is restricted by the homogeneity of politicians, and random selection would help tap into the "wisdom of the crowd".
Carson says even representatives who don't know much have an important role to play.
"They'll be asking really naive questions or playing the devil's advocate," she says. "It's all fodder for deliberation."