It has long been my thesis that as data collection becomes more comprehensive and provides easier analysis for things like credit risk, criminal prevalence, diseases, fraud rates, etc. that the picture will emerge of subgroups in the population that are so different from the norm - which foist such massive costs on the rest of the population - that when the full magnitude of it is known there will be political support for efforts to allow discrimination in matters like auto insurance, etc.
If for instance, it was known that loss rates for one ethnic group of drivers were in the same proportion as the HIV+ rates quoted above (58.2/2.9=20x) would people support a surcharge on a group which cost an insurance program 20x as much as the rest of that segment? I think so. The reason that there hasn't been support for this to date is that the information has not been available. I'm sure insurance companies run and hide from the issue - but at some point they may be required to publicly disclose the segments of the insured population that are driving costs out of control. It may not be segmentation based on race - but with numbers like above, I cannot image a segmentation scheme that would drive the numbers any higher, except maybe geographic sub segments (e.g., black women in the SF area - has to be a shocking rate).
What's the rate of IV drug use in these groups?
Every inidividual is an individual, Wally. I've had enough in my life of segregation, of racial stereotyping. I grew up in a segregated America -- it sucks. I like to take each person for what he or she is, individually.