Six Lines

Data Mining for Life Insurance

Posted by Aaron Massey on 19 Nov 2010.

The Wall Street Journal examines the use of data mining techniques to calculate risk for life insurance companies. Although the article is good, the idea and the technology behind it is nothing new. One of the books that got me interested in security and privacy when I was an undergrad was Database Nation, which discusses this concept at length. I suspect it will only become more common, and I’m not completely convinced it is a bad thing.

Serious privacy concerns exist as a part of any mass data mining of consumer habits. No one wants to have every purchasing decision they have ever made put into a database and analyzed for eternity. This is particularly true for some sensitive purchases; some aspects of our consumerism must remain completely private.

Consumers also stand to benefit from data mining. Insurance is meant to hedge against uncertainty. It’s all about knowing that you’ll be financially covered even if you get hit by lightening or randomly attacked. Insurance gets cheaper the more “certain” we can be. If it turns out that some purchases really do indicate a risky lifestyle, then choosing to purchase them anyway is an act of personal responsibility. Similarly, everyone knows that regular exercise is important to living a longer, healthier life. If your purchasing history details how committed you are to exercising and you want cheap insurance for unforeseeable possibilities, then you may want this sort of comprehensive data mining.