Significantly, that’s not counting the tsunami. AIR’s press release points out that you can’t just add a tsunami estimate to its quake estimate. That would be double-counting a lot of buildings that were destroyed by the quake, then swamped by the tsunami.
A tsunami estimate is in the works, the company indicates.
AIR says it has been hampered because Japan’s seismic network, K-NET, is offline.
So AIR had to estimate the quake’s rupture dimensions – the length and depth of the section of fault that slipped:
. . . AIR simulated dozens of scenarios with varying magnitude (8.9 to 9.1), focal depth (15 km to 30 km) and rupture width (100 km to 150 km). The losses are most sensitive to rupture dimensions, and become extremely large if the modeled rupture is extended southward towards the Tokyo and Chiba prefectures, which contain a higher concentration of insured properties.
If the rupture extended south, toward Tokyo, the estimate trends higher. Early reports indicate the rupture extended the other way.
The estimate also doesn’t count demand surge, the phenomenon in which the cost of construction labor and materials rises because so many people need contractors at the same time. AIR notes there was not much demand surge after the 1995 Kobe earthquake (also called the Great Hanshin Earthquake). A repeat of the Kobe quake would cost the industry $7 billion.
AIR’s estimate is considerably larger than the first numbers to emerge Friday, which were on the order of $10 billion.
AIR’s top-end number is about the same as its estimate for a repeat of Great Kanto Earthquake of 1923 ($35B).
Lots of additional information at the link.