Here, for instance, is a chart of H1 natural disasters in the US going back to 1980. While the number of geophysical disasters has remained very low, the number of climatological disasters has more than doubled. And it's still growing fast: the number of disasters in just the first half of 2008 was bigger than the number of disasters in any of the full years between 2001 and 2005.
The global trend is also clear: Of course it's not just the number of disasters which is increasing: it's also their severity, and the insured losses associated with them. Here's what's happened to US thunderstorm losses in the first half of the year, going back to 1980:
These losses, it should be noted, do not include any damages from the devastating midwest flooding in June: floods and thunderstorms are two different things, for insurance purposes.
There are similar charts available for wildfires and winter storms too, but clearly by far the biggest natural-disaster risk facing the US is hurricanes, which over the long term have historically accounted for roughly half of total insured losses. It's hard to find a climate scientist who'll claim that the risk of severe hurricanes isn't growing - even as the insured value of property on the US coast is rising fast. Insured coastal exposure in Florida, for instance, was $1.9 trillion in 2004; by 2007, that number had risen to $2.5 trillion, with total coastal exposure in the US reaching $8.9 trillion, largely as a result of the boom in construction activity.
These are enormous numbers, and help put the insurance industry's policyholder surplus of just over $500 billion into some perspective. Yes, that's an enormous amount of money. But if a hurricane hit Miami, or Houston, or New York, it's not hard to imagine how quickly the bills could add up - bills which would be paid overwhelmingly by private insurers.
Does it even make sense for private insurers to take on that kind of hurricane risk? I'm not sure. Certainly attempts to offload the risk onto the capital markets, using catastrophe bonds, seem to be going nowhere any time soon: total cat-bond issuance was less than $8 billion in 2007, and that was by far a record year. But if private insurers are to take on this kind of risk, then I can certainly understand why they'd want to raise insurance premiums in anticipation of increasing hurricane frequency and severity.
Homeowners won't like it, and regulators won't like it. But increasingly-severe natural disasters are an inevitability at this point, and they'll hit the US just as much as they'll hit the rest of the world. In other respects, the US is actually in a fortunate position with respect to global warming: other, poorer countries are at much greater immediate risk. But when it comes to potential insured losses from hurricanes, the US is by far the world leader.