Mathematicians: I predict a riot; where the next dictator will fall.
by New Scientist:
Scientists who study mathematically complex systems claim we can do better. They are planning to study recent events to devise better ways to predict a fall.
Complex systems with many interrelated variables, such as ecosystems or societies, can accumulate stresses while showing no obvious change – until they reach a point where a small stress can trigger a sudden shift to another stable state. For example, forests accumulate kindling until a spark ignites a fire.
According to Yaneer Bar-Yam, who heads the New England Complex Systems Institute in Cambridge, Massachusetts, the stresses of poverty, unemployment and an absence of government accountability built up in Middle Eastern countries with a large “youth bulge” of young adults without jobs, children or prospects. Then spiking food prices and the public suicide of one young Tunisian triggered revolution.
The key to predicting regime shifts, says Marten Scheffer of the University of Wageningen in the Netherlands, is to look beyond individual behaviour to seek simple laws that describe a population’s collective behaviour.
Bar-Yam has previously used mathematical models to predict violence between ethnic groups. Though the system’s mathematics was complex, it yielded a simple result: ethnic violence flares when enclaves are a certain size. This successfully modelled 90 per cent of recent ethnic conflicts in India, Kenya, central Asia and former Yugoslavia (Science, DOI: 10:1126/science.1142734). With the right data we can model other social changes, he says – though good social data may be hard to find.
Scheffer, however, believes such data may not be necessary. “All complex systems exhibit certain symptoms before a regime shift,” he says, including slower responses to small changes, and a tendency for all players to behave similarly. Bar-Yam has found this behaviour pattern in the lead-up to market crashes. Scheffer is launching research to look for such symptoms in social systems, including the Middle East.
In the past, Scheffer says, analysts focused on the trigger that sparks change, rather than the underlying system. “We cannot predict the spark,” he says, “but we can say when a forest has accumulated dangerous levels of kindling.” Repressing revolution is not the way to achieve stability, he adds. It would be like preventing small forest fires, allowing kindling to accumulate until a big fire breaks out. But uncovering the symptoms of instability may warn societies to reform themselves before revolution happens.
Why isn’t this being taught across universities? A course teaching the statistics of social unrest could prove to be valuable for future human statisticians or actuaries.