Predictive Models Are Wrong. I Yawn in Rage.

Not sure if you folks noticed, but much of the catastrophe reinsurance world is stammerin’ mad.

To understand why, it’s important to know how one starts a company specializing in catastrophe reinsurance:

1. Rent an office (as in ‘room with a phone and internet connection’, not ‘floor in midtown Manhattan’. It’s mostly for mail anyway)

2. License a catastrophe model (choose from one, two, three)

3. Have a billion dollars

4. Move to Bermuda

Notice the lack of administrative staff, assistants, sales team, etc. The reason for this is that everything is outsourced.

Everything, that is, except timing.

You see, these companies make a lot of money when there are no catastrophes. They lose a lot when there are catastrophes.

All you’ve got to do is talk someone into giving you the cash. I liken it to buying a gun (in the US).

Sure, you have to more or less prove you’re not a homicidal maniac to get one, but any moron knows what to do with it if his life is in danger.

Same with cat reinsurance lately. In about 4 of the last 10 years, any complete idiot would have made a fortune if entrusted with a billion dollars. Some did.

In about 3 of the 10, the savviest, most epic douchebag ‘genius with ADD’ would have lost a fortune. The other years are for the pros, who genuinely can differentiate themselves.

I tell this story because a very important part of this picture is the catastrophe model.

The advent of the cat model (in the early 90s) changed everything. It commoditized cat reinsurance. This coincided with a dramatic increase in cat rates, as it happens, but it also coincided with a dramatic increase in cats.

Cat models are a map with a hurricane/earthquake simulation. You load the coordinates of the buildings you insure and the model pretends to smoke them with natural disasters. It lets you diversify the portfolio among independent regions (European Wind and US Earthquake, for example).

The lazy among us genuinely believe my little list of four above and don’t really understand the models they use. These are the people that get royally pissed off when the cat models change.

Which they are doing now.

I won’t get into the details, but I’ve had to bite my tongue a few times listening to people whine about how ‘models are always wrong’ and how they want to ‘sue RMS’ or some such nonsense. I want to shake these people.

Of course models are wrong. If you know so damn much, why don’t you build a model that says different.

Advertisements
This entry was posted in insurance, insurance business strategy. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s