The main problem I see is the adverse-selection problem: if it's not via a large group pool like a large company, then there is likely to be significant correlation between poor health and higher insurance-purchase rates, which will in turn drive up prices. In the individual market, if I'm healthy vs. if I fear I will have imminent health problems, I'm more likely to buy insurance in the 2nd case. But if you work for IBM, everyone gets insurance, uncorrelated with their health perceptions--- so insuring IBM is probably a better deal overall, because the risk pool doesn't select for worse risks.