In an interview last week with Simon Murray, he cast doubts on the validity of economic forecasts because he said everything was dependent on political events, and cited acts of war as chief among them - ie if Bush declares war on Iraq tomorrow, all forecasts have to be recast.
Murray's point is an interesting one, and given current circumstances with the War on Terror, led me to do some research about war and its frequency. We are accustomed nowadays to think wars are rare events, and this is possibly why financial markets react so strongly to them.
Actually, if you closely analyse a historical sample, you see that wars are not rare events...