Why are some crises contagious while others die at the border? The explosions during the 90s in Mexico, Thailand and Russia were felt across the world. The next wave in Brazil, Turkey and Argentina, on the other hand, barely roused investors off their dotcom bean bags. This latest meltdown, of course, was financial smallpox.

Economists have tried to explain different levels of contagion over the years, from theories about trade and the similarities between countries, to reactions by guardians of capital such as banks or mutual funds. There is some consensus around the idea that the level of anticipation of the initial crisis makes all the difference. Looking at indicators such as bond spreads or fund flows, it seems when investors see the big one coming it does not spread.

But why? A new paper from the International Monetary Fund reckons uncertainty is the key. When a crisis occurs in one country, investors' belief in their intelligence-gathering ability at home diminishes. Rattled, they adjust their behaviour, which in turn increases spreads, thereby precipitating a new crisis. Furthermore, when meltdowns abroad are anticipated, the level of uncertainty in an investor's own country actually rises, as if to say: “Well I guessed that mess was going to happen – my forecasts for my own market must be spot on.”

The policy implications of uncertainty as the root of contagion are obvious. Try to get rid of the shocks. Regulators must wave the flag early. More important, because regulators are prone to miss elephants in living rooms, get as much detailed economic and financial data into the public domain as possible and let investors work it out. If, for example, China blows up, it won't be half as bad for the rest of the world if the numbers give the game away before hand. Worst would be some official admitting that the growth data was bunk all along.