Thinking through comparisons on Medicare

This weekend I was listening to Kathleen Sebelius testify about Medicare. One of the things that came up in the hearing was a comparison of Medicare payments to private payments. I forgot the figure exactly, but Medicare was supposed to have compensated physicians something like 80% of the amount that private plans do. I'm no health care expert but I believe this is different from the co-pay (which I understand is also 20%) because in the discussion they were actually talking about privately insured patients cross-subsidizing Medicare patients.

I had much the same uncertainty about this discussion that I did in the post where I caution against making head-on comparisons of public and private school students. How do they make this comparison with Medicare? For one thing, you have to be comparing seniors to seniors, controlling for their conditions, etc. But that's what introduces the problem. The large majority of seniors are on Medicare in the first place. Those that use private insurance are qualitatively different sorts of patients than Medicare patients. A straight comparison of the two would seem to be biased against Medicare because the mere participation in a private plan by a senior in the United States indicates a higher willingness to pay. Another similarity between Medicare and my public school example in the link above is that without Medicare, uninsurance rates for seniors would probably be comparable to the rest of the country (if not higher). So what is the right margin of comparison? What is the right counterfactual to think about? Seniors on Medicare now compared to a self-selected group on private insurance now? I think the more reasonable comparison would be everybody now compared to everybody in the absence of Medicare. If we think of how Medicare impacts costs and physician reimbursement, that's the right comparison. If we make that comparison, these numbers are likely to look a lot different. More senior citizens would be utilizing emergency rooms and charity care and more would be uncovered.

Either way, it's not clear to me (1.) how they get the underpayment comparison with private plans without a selection bias problem, and (2.) whether that's even the right comparison to be thinking about.

A lot of the points made in the hearing about the Ryan plan made sense (and have always made sense) to me. One of the concerns about Medicare Advantage (the current "private" version of Medicare) has been that it is subsidized, so a head-on comparison with regular Medicare is hard to make. The Ryan plan doesn't do this, of course, so it avoids that problem. I would support a voucher program as an option for Medicare, at least until we know more about how it works. Again, you're going to have a different group of people selecting into such a voucher program. What I would do is give all Medicare beneficiaries the option of converting their benefit into a voucher, and then randomly assigning who actually gets the voucher from among those who choose it (i.e. - if one million Medicare beneficiaries choose the voucher route, randomly select 500,000 to actually receive it so that we can subsequently do comparisons). There are good reasons to think that vouchers would solve a lot of the problems with fee for service, which is good because nobody likes the fee for service system.