There always is some form of subsidy in insurance that doesn’t measure real risk, health insurance has been one of the few that basically incorporates risk into the entire pool (so for example, assuming that 20% of the population smokes, incorporating that into the entire pool as a risk, and calculating rates based on that, same for things like age (since older people tend to use health insurance more), and the like. It is one argument for a single payer system, that with a gigantic pool that the risks are spread out, and that younger people help subsidize those older.
Life insurance, because it is direct, takes all the factors into consideration, if you smoke, are overweight, have a family history of heart disease, cancer, etc, it all factors into your cost of life insurance.
Other forms of insurance do shift costs from Peter to Paul, flood insurance is a classic example. Where I live is maybe a 1 in a hundred year flood area, no big rivers near by, no lakes or ocean, yet my flood insurance is a significant percentage of my total house insurance (it is around 600 a year for the flood alone). Why? Very simply, that attempts to make flood prone areas have to pay the full risk of where they live, has run into politics. If it weren’t for this cost shifting, people in flood areas near the coasts, in areas in the midwest near rivers, and so forth, would be paying significantly more to get flood insurance (some places that are extremely risky, like barrier islands, I believe cannot get it) and it might as well be unaffordable (an article said that for example, a homeowner near the water in Queens, NYC, could see his flood insurance go from 500 bucks annually (which is less than I pay!), to as much as 15k a year. The flood insurance program is in a financial hole in part because of the cost of major storms, especially Sandy, hit it hard.
With auto insurance, people living in suburban and rural areas often pay a lot more than they otherwise would based in risk, because insurers are not allowed to charge the real rate of rick in city areas and known dangerous places to drive, because the insurance could become unaffordable to people, and given that in many places you have to drive to get to work, would be a major problem.
I think one of the problems with smoking on health insurance plans is that given the huge costs associated with smoking related health problems, incentives to quit smoking (like cutting the rate a person pays when they quit smoking) may not be incentive enough, if let’s say they estimate the risk of smoking is worth 200 bucks a month in risk, it wouldn’t be fair to drop a smoker’s rate 200 bucks a month so they end up paying less than someone who never smoked, whereas charging a smoker 200 more a month than dropping it 200 if they quit is fair to everyone else. With smoking it is huge hanging fruit,other than a few people who might tell you “nothing wrong with smoking, grandpa frank smoked until the day he died at 96”, almost no one denies that smoking as a health risk outstrips almost anything else, is so closely correlated to costs that it can’t be denied. Yep, the risk of getting let’s say type II diabetes increases with weight, age and how you carry the weight, but that risk pales in comparison to smoking related costs/risks (so if there is a 20% higher likelyhood of getting type II diabetes if you are overweight, that means that significant proportions of those who are overweight will not get the disease, same with other ills of weight)…obviously, that also changes if someone is morbidly obese, but the other thing with weight is that the old puritanical notions that it is always a moral failing have fallen by the way side, it is a fact that genetics plays a role, that the skinny fashion model who eats junk food and swills soda and is a size 0 is freakish genetics, and that the person who is morbidly obese might eat carefully, and still be way overweight…and with genetic related causes, it becomes a lot harder to compare those to someone smoking, which basically is addiction and something they chose to do.