Thursday, July 3rd, 2008
warming up the models

In a front-page story on Tuesday, the Wall Street Journal examined the relationship between rising sea temperatures and rising insurance premiums, with a special focus on the mechanism that drives the linkage — catastrophe modeling (so-called “cat models”).The Wall Street Journal | This is a good introductory article about cat models, although their mind-numbing complexity can only be hinted at, even in a longer piece like this. The political implications of higher insurance premiums and homeowners with catastrophic losses are much more predictable. No matter your “take” on global warming (and in the opinion section of the same issue of the Journal, Bret Stephens left no doubt about hisThe Wall Street Journal | Stephens calls global warming a “mass hysteria phenomenon” of “sick-souled” religious zealotry.), the prospects of it provide a wealth of considerations for investors.

Let’s say we knew for sure that global warming was real and that we could reliably plot its impact over the next fifty years. Certainly its effects should be included in our models for all manner of companies and investment products. On the other hand, if we knew that the climate would vary in a random way no matter what mankind did or didn’t do, we could comfortably leave it out. Thus the boundaries of our set of possibilities happen to coincide roughly with the dogma of the believers on one end and the disbelievers like Stephens on the other.

If we are to remain coolly analytical, we must concede that we don’t know anything for sure, and take a probabilistic approach. Because they see only certainty, those without doubt (on either end of the spectrum) struggle to apply the simple yet powerful concept of expected value, where the possible outcomes are weighted according to their probabilities, yielding through basic mathematics an average expected “payoff.” Those schooled in popular culture will recognize this process as the function of “the Banker” on the TV show Deal or No Deal — he’s just an expected value calculator behind a darkened window.

Since the debate on global climate change tends to be more theological than anything else, perhaps “Pascal’s Wager”Stanford Encyclopedia of Philosophy | To quote this source, “rationality requires you to perform the act of maximum expected utility. Therefore, rationality requires you to wager for God.” is the more apt reference; it also takes us back to the very beginning of probability theory. Even if the odds of profound effects from climate change are quite small, if the effects are, in fact, profound, they can have a sizable impact on the analysis.

But enough theory and philosophy. The Journal article sets up a number of important questions:

What is the “right” history on which to base an analysis — the last five years, the last hundred years, or what? (Investors are often told to start with the magical year of 1926.)

Should the expectations used in the model always be based on that history, or should they be adjusted in some manner for today’s state of things? In the article, this debate boils down to how to respond analytically to a belief that our current warmer ocean temperatures bring a higher likelihood of hurricanes for some time to come, not dissimilar from the choice investors faced at the turn of the century when P/Es on stocks were in the 30s and 40s.

Institutionally, how do we avoid the tendency noted by Karen Clark, a cat-model pioneer, to use our models “as if they produce definitive answers rather than uncertain estimates”?

And finally (at least for this posting), will the models of analysts that focus on companies start to include the effects of global warming — or the effects of anticipated governmental, commercial, and personal actions to lower the probability of global warming — sometime soon? If so, will the pricing of securities change as a result?

You might want to think about it before the Banker gives you a price in front of a screaming crowd.