Friday, June 8, 2012

Political Process of Negotiating Hurricane Risk

Several days ago, the South Carolina Post and Courier published an article by Tony Bratelme.  The article clearly describes the existing disagreement about the true measure of the hurricane risk at the heart of the windstorm insurance problem that several states are experiencing.  Bratelme's article demonstrates that characterizing the hurricane risk is a political process that rests on interpretation and valuation of science information.  In the end, the cost of insurance reflects the political risk that is created in the process of negotiating agreement about the hurricane risk.

The article follows the travels of a South Carolina resident who seeks to identify the "true" hurricane risk.  The resident eventually finds, of course, that there is no true hurricane risk.  There are any number of statistical measures of the risk and their is everyone's chosen valuation and importance of such analyses but, there is no one true risk measure.

Bratelme first presents a basic analysis of the historical record and then places a valuation on the numbers:
Since 1851, 30 hurricanes had spun within 50 miles of South Carolina. That was one about every five years, which seemed to be the definition of vulnerable.
As he digs deeper he decides that the data suggests that the risk is minimal but

that didn’t really say much about the long-term vulnerability of a particular house in Beaufort or Charleston.
So, he looked to the now common practice of catastrophe modeling.  Upon the Post and Courier's request, prominent tropical meteorologist Kerry Emanuel produced his own statistical analysis with models he uses in his company, WindRisk Tech LLC.  With Emanual's modeled risk in hand, Bratelme determined that the risk is lower than residents are led to believe,
It meant that the truly disastrous ones [hurricanes] were rare and certainly not the impending train wrecks that weather channels and emergency planners sometimes suggest are on the way.
To others however, the same data does indeed suggest an impending train wreck.  For insurance companies, severe hurricanes, although rare, can lead to economic ruin.  Experience with Hurricane Hugo in 1989 and Hurricane Andrew in 1992 heightened insurer perception of "catastrophe" risk.  Such shocks can have grave impacts on the larger socioeconomic system.  The concentration of wealth along coastlines is a concern for insurers even if the probability of loss is "low."

Methodology can incorporate additional risk perceptions in the characterization process.  Bratelme identifies one (of many) instances where this has occurred.  Emanuel and his colleague Michael E Mann, theorize that a lull in North Atlantic basin hurricane behavior and its recent heightened activity is due to pollution production.  Therefore, reflecting on pollution levels to calculate hurricane risk estimates to then calculate insurance premiums incorporates elements of environmental risk (ie. changes in pollution levels) into the price of the policy.  

In the process of writing his article, Bratelme identifies a new dimension of the hurricane risk that is developing.  There seems to be a growing feeling that the means of characterizing the hurricane risk through the use of catastrophe models is a threat (see here too).  Recently, the Massachusetts insurance regulator turned down requested rate hikes specifically citing the models as a reason for doing so. (But of course, such a decision influences insurers' fears of insolvency)

And so, Bratelme nicely shows that while there are many statistical evaluations of the probability and severity of loss, there is no one true hurricane risk.  Likewise, there is no right estimate of the risk.  There are instead, estimates that are better suited for some interests than others and model estimates can be used as tools to support one's chosen risk perspective.  The cost of windstorm insurance reflects a political risk that is created as a result of negotiating compromise between interests and their chosen risk estimates.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.