LOTTERIES AS EFFECTIVE DECISION-MAKING PROCEDURE: application of Ancient Athenian sortition
03 May 2019
Using lotteries can be as effective as democracy in making collective decisions, according to research which considers the Ancient Athenians’ use of sortition to select public officials.
This study by Rene Saran and Norovsambuu Tumennasan, published in the May 2019 issue of The Economic Journal, finds that while sortition cannot guarantee political equality, the process of selecting decision makers by a lottery is a potent alternative to organising elections or referendums. They also argue that sortition can help improve decision making in business organisations.
In democratic societies, individuals give their preferences by either voting directly for the policies or for the representatives who subsequently decide on policies. However, both require substantial investment of time and money so that each eligible voter has the opportunity to vote. For example, it has been estimated that waiting in line during the 2012 US presidential election cost around $544.4 million in lost earnings.
The authors consider the process used by Ancient Athenians to select several public officials and priests by the drawing of lots (sortition), which is inexpensive compared to elections or referendums. However, it is rarely used today to make collective decisions except for selecting juries.
Modern-day democracies treat all citizens equally, giving each individual one vote. In contrast, citizens do not have equal say under sortition. Even if every citizen has equal probability of being selected, those not selected have no say at any stage of the decision-making process.
The researchers focus on environments where the collective decision depends on information that is dispersed among interest groups. They identify effective sortition procedures by two properties:
- Either there is a single individual in each interest group (e.g., the group leader) who is always selected, or every selection includes two or more individuals from each interest group.
- No interest group is partitioned into disconnected subgroups in the sense that individuals belonging to distinct subgroups are never selected together.
Examples of effective sortition procedures include an opinion poll, where a random sample of two (or more) individuals from each interest group is selected, and “oligarchy with random audit", where the group leaders are audited by a randomly selected individual from the rest of the population.
Apart from selecting a small number of decision-makers using lotteries, the effective sortition procedures differ from democracy in another fundamental way: Instead of voting, individuals report their information and make contingent bets regarding the reports of other individuals (like in a predictions market). When accompanied with carefully defined fines if someone is caught lying, these sortition procedures guarantee truth telling by all.
The authors argue that this study also contributes to decision-making in businesses. Business decisions in large corporations depend on information that is diffused across the organisation. In traditional hierarchical organisations, lower-level employees do not communicate with upper-level management. For instance, the President of a company will usually make strategic plans while consulting only the Vice Presidents of different departments in the company, who in turn consult employees in their own departments.
This study shows how adding sortition to such traditional business hierarches (e.g., by selecting lower-level employees to randomly audit the Vice Presidents) can improve decision making in the organization. These results make a strong case for corporations to adopt comprehensive whistleblower policies that encourage and protect whistleblowers.
Eliciting and Aggregating Information By Sortition in Collective Choice by Rene Saran and Norovsambuu Tumennasan is published in the May 2019 issue of The Economic Journal.
Associate Professor | University of Cincinnati
Associate Professor | Dalhousie University