|
 There have been both more frequent and more severe
catastrophe losses in the last ten years than in the
preceding decades.
Some U.S. politicans and some large reinsurers have
associated these increases with global warming and
specifically with man-made changes in the
concentrations of “greenhouse gasses”
Since both parties are selling something it is worth
undertaking an independent review.
The best U.S. record is the PCS database.
PCS Data – All Events, Total by Year
But is a 1949 catastrophe the same thing as
a 1998 catastrophe?
PCS data – changing definition of a catastrophe
Dollar thresholds that change over time
Dollar thresholds that can’t directly reflect inflation,
population density and increases in real insured values
Changing interpretations of “widespread”
Improved communications and access to remote
locations allows reporting of more events
Adjustments to the catastrophe data are needed
Past losses brought to 1998 levels using the U.S. GDP as a trend.
- Explictly adjusts for population and price level increases
- Implictly adjusts for increased real insured values
After this adjustment we can see the difference in reporting
criteria, as some years have many relatively small losses and
others have none.
- We excluded all losses below $45,000,000 on a 1998 level
- Selected to equate the on-level value of the reporting criteria
for each year
Analyzed different catastrophe perils separately.
- Presumably, CO2 levels don’t cause earthquakes
- Hurricane vs. other wind vs. EQ vs. all other
|
