It is highly likely that the insurance industry will soon face similar regulations to those already faced by the banking industry (Mertz 1993a). The reporting is likely to be at the zipcode (postcode) level. However, since this is in the future, this section on insurance will focus on current applications: risk assessment and avoidance. In the US, natural disaster after natural disaster has occurred over the last four years, bringing the total insurance industry’s bill to $34 billion (Mertz 1993b). Because of this, the industry has begun seriously to consider how it underwrites policies. The Oakland fire, Hurricane Andrew and the floods and other catastrophes of the summer of 1003 each pointed out the need for greater underwriting care. Insurance companies are beginning to realize that they must either refuse to insure properties that are at great risk from natural disasters, or justify larger premiums for doing so, or clearly identify precise areas in which property damage may be partially reimbursed from other sources. For example, the cross-hatched areas in the figure are wind-pools: insurers who write property-damage policies in these areas may be partially reimbursed by a state fund.
Figure: GIS for insurance risk assessment. Caloosahatchee River, US (grey: low risk; hatched: high risk)
Analysing properties in relation to the geographical likelihood that they will be damaged because of a natural disaster is not possible without using geographical analysis (Runnels 1993). These kinds of application are good examples of “embedding” because they primarily depend on a property being compared with a “geo-file”, or a listing of addresses that has been coded for likelihood of damage using GIS. In this instance, operators are unlikely to see a map during the underwriting process.