A year after the implosion of Lehman Brothers sent world markets into turmoil, the question of where the next global shock will come from - and whether it can be predicted and prepared for - has never been so urgent.

What makes the issue particularly difficult is that many of the events catastrophic enough to cause a major crisis - known as "fat tail risks" or as "Black Swans" by trader and author Nassim Nicholas Taleb - come from outside the realm of finance.

To be able to forecast what the next global shock will be, we need to be able to make predictions about geopolitics, war, terrorism, extreme weather events, earthquakes and pandemics.

In their book this year on fat tail risks, Ian Bremmer and Preston Keat of political risk consultancy Eurasia Group noted that they pose fundamental problems for accurate prediction.

Fat tails, they wrote, "represent the risk that a particular event will occur that appears so catastrophically damaging, unlikely to happen, and difficult to predict, that many of us choose simply to ignore it. Until it happens."

A growing body of theory and evidence suggests that making accurate forecasts about rare catastrophic events is inherently impossible. But it also suggests a practical solution to mitigate the dangers - detailed scenario-planning by imaginative analysts who do not cling too tightly to mathematical models of reality. Whatever analysts attempt to forecast - the economy, the weather, the progress of epidemics, geopolitical change - the key problems are the same. Systematic forecasting requires a model that approximates reality. But is this feasible?

Many models assume simple linear relationships between variables, but there are plenty of reasons to believe that in the real world, variables often react in a complex, volatile and non-linear way, particularly where extreme events are concerned.

Models also rely on the future resembling the past. But Mr Taleb argues that by their very nature, "Black swan" events lie outside of normal experience and up-end traditional assumptions.

An even more fundamental spanner in the machinery of many models is that they need to find a way to capture the behaviour of a particularly volatile and unpredictable element - us.

Economists and political scientists have long worked on the assumption that people largely behave in a rational way that can be modelled and predicted. That assumption has been left in tatters by the global meltdown of the last couple of years.

Globalisation has meant forecasts now need to take into account hugely complex human interactions among millions or billions of people. And forecasting everything from war to weather has been further complicated by the fact that humans increasingly affect the environment around us in profound ways.

Empirical evidence on the accuracy of political forecasting is - to say the least - not very encouraging.

In the most ambitious attempt to analyse the effectiveness of political forecasters, psychologist Philip Tetlock polled 284 political and economic experts for multiple predictions over a 20-year period, for a total of 82,631 forecasts. The overall result was that expert opinion did no better than pure chance.

"When I have staged competitions, many forecasters fail to outperform the proverbial dart-throwing chimpanzee," he wrote in a review last month.

So where does this leave investors and businesses? Should we just accept we are helpless in the face of unquantifiable risks that we cannot foresee - what Bremmer and Keat call the "We are all doomed" approach to risk management?

Not necessarily. Many analysts say scenario planning and risk mapping provides at least a partial solution. While we may never be able to predict extreme events with any certainty, investors and executives can prepare for the future by analysing worst-case scenarios and considering how they would deal with them.

"Scenario analysis is a particularly useful means of understanding uncertainty and fat tails," said Bremmer and Keat. "The main objective is to inspire creative problem solving and to spur managers to think about unthinkable outcomes."