Economists like to distinguish between "thick" and "thin" markets. The quick distinction is that thick markets have lots of buyers and sellers and high volumes, meaning that conditions in the market are likely to be closer to the textbook description of pure competition. In a purely competitive market, profits are dissipated, market clearing prices are stable, and production occurs at marginal cost. In a thin market, there are fewer buyers and/or sellers, lighter volume and the potential for monopoly rents. A typical example of a thick market would be spot markets in grain commodities: something like 20 billion bushels of corn are produced annually worldwide, much of which is traded in the CBOT. A thin market would be custom yachts or original fine art. Livestock auctions are thick; Sotheby's auctions are thin.
Unfortunately, something is missing from this standard description. Economists like to think of thick markets as efficient and it's a little embarrassing when efficient markets go awry. One way to save face is to invoke probability. If future events are not well-ordered (think Taleb's Black Swan or North's non-ergodicity arguments) or predictable according to a known probability distribution, then systematic bias has fertile soil. Seed that with some common cognitive errors and plow it with distorted nominal prices thanks to wobbly monetary policy and hey-presto, reap widespread market instability with or without piecemeal policy tinkering by the legislature.
So what's the connection between economic crisis and insurance? Plenty. Insurance is a way for people to hedge against disaster. I buy car insurance because, as an individual, I bear some risk for wrecking every time I get behind the wheel. I buy life insurance because I could get hit by a bus crossing the street. I buy homeowner's insurance in case... well, I'm not 100% sure offhand what's covered in my policy, but you get the idea: there are predictable risks out there and even if I don't know my own individual probability down to the fraction of a percent, when I'm put into an actuarial pool of people mostly like me, my probability can be reasonably simulated. That way, as long as the market is thick, we can get reliable prices for different insurance products. In the aggregate, most of the above-board insurance policies you can get (whole life, auto, home, renter's, et al) are euvoluntary.
Some aren't. Think of the sleazy used car salesman with the "so I'm sure you want the extended warranty, right?" line. Think of sketchy overdraft protection you can buy for some bank accounts. Think of shady ticket sales that have you paying out the nose for the privilege of canceling without heavy penalties. Some insurance is generally seen as wise and prudent, other as a sucker's bet or exploitative. These fools' policies could hinge on any one of the ingredients I listed above on the financial crisis farm: unknown probability distributions, cognitive error (or asymmetric information if you prefer), or distortions in interest rates. Usually however, they come smack-dab face-to-face with ex post regret. Sooner or later, so the assumption goes, the poor fellow who bought the extended warranty will look back on the folly of his error and be left feeling ripped off. Poor fellow. Poor, poor fellow betrayed by his own foolishness in the thin market.
Well, that's easy to say after the fact. The problem with Cassandra is that before Agamemnon lies dead, she's just another crazy person jabbering about the future. There's no way to distinguish her from false soothsayers. Ditto for people who wrung their hands over housing prices in 2007, ditto for Ralphie's mom chanting, "you'll shoot your eye out" and ditto for the well-intentioned paternalist clucking a manicured tongue at the rube signing up for a turtle invasion policy. Peering into the future is hard enough, peering into the minds of others as they peer into the future is impossible with current technology.
Claims that odd, thin or irregular forms of insurance demand regulation or prohibition require a substantial burden of proof. Evidence of coercion (by human agency) is sufficient, but what else might cut the mustard? What would you consider to be a sufficient burden of proof for you to interfere with a voluntary insurance sale?
Bonus question: if you squint your eyes right, insurance can be seen as a form of gambling. Do the same arguments that apply to weird or suspect insurance plans also apply to gambling?