You have the correct thesis, I think, but some poor examples:
Pollution - everyone bears the cost but only the polluter benefits.
Not exactly; the people who pay low cost for the goods/services provided by the polluter/overfisher also benefit.
The examples of credit cards are better. MBSs, I'm not so sure - I don't think there was "tricking" there so much as an artifact of booking rules associated with unrealized gains and losses. Add to that the hot-potato nature of financial instruments, and it's unsurprising, really.
Insurance is a different beast - I've come to realize that insurance isn't about reducing total societal costs at all but it really is about socializing the cost. So large pools of people pay a total higher cost for potentially lower out-of-pocket costs per individual. Insurance by itself cannot ever reduce costs, especially health care costs - the providers have all the power, because sick people are basically willing to pay whatever it costs to be made well.
So the only way health care costs can come down is if people are more healthy, there are more providers, or there is regulation imposed to cap prices (e.g., all the recent hoopla about pharma companies buying rights to drugs and increasing prices because "that's what the market will bear").
The current regulatory framework almost guarantees fewer - not more - providers, so there is really very little hope for reduced health care costs from that aspect. So the only hope is that the socialized preventive care really does effect the desired increase in overall population health - but we won't know that for probably a decade or two.