I see a few candidate hypotheses:
1. Adverse selection in insurance markets. In this view, outside of the corporate employment context, mostly the unhealthy — the "lemons" — show up to buy health insurance.
2. Poor information about the cost and benefit of different medical procedures and providers. I am puzzled why we don’t have better institutions for evaluating providers and spreading this information to consumers. Part of the problem is legal, part of the problem is measurement (doctors could dump near-hopeless patients to get better ratings), and part of the problem is that many people don’t want to know how good (read: bad) their doctor is.
3. Time consistency. Ex ante, we are most worried about catastrophic health risk. Ex post, most of the overinvestment in health care comes for victims of catastrophic health risk. No set of institutions can square these dual perspectives satisfactorily.
4. Showing that you care. This is Robin Hanson’s hypothesis. You spend money senselessly on health care, mostly to show your wife you love her and the kids.
Believers in national health insurance tend to emphasize number one, which might be alleviated by forced mass pooling. In contrast, I am skeptical that adverse selection is the significant problem.
It is less clear that national health insurance could improve performance on number two. People would remain underinformed. Government might have less incentive to rip them off, but the implication would be that government provision is "lazier" in general. Not a comforting thought.
National health insurance does address numbers three and four, albeit in backhanded fashion. Ex ante, people feel protected and the program is popular. Ex post, such systems spend less money on the last six months of life than does the U.S. system. The relevant denial of treatment is often invisible rather than a stern hand pulling the plug.
The bottom line: Defenders of NHI place great stock on #1. If #1 were significant, we could, at least in principle, use national health care to both lower costs and improve treatment.
If #3 and #4 are the major problems, national health insurance provides benefits by restricting overinvestments in health care. This is consistent with Europeans living longer and spending less on health care. The U.S. maybe could replicate these benefits if a) we push out private insurance companies, b) we ration or abandon expensive procedures which don’t extend lives very much, and c) we adopt healthier lifestyles.
I am skeptical of #1, and I am not ready to bet on a), b), or c), much less all three at once.