Extended Discussion Of Lewin Group

Here’s a more technical discussion of the Lewin Group’s new report on public plan choice for those who can’t bear to miss the trees amid the forest.

First the briefest possible background: Public plan choice means that people who don’t get coverage from their employer get the choice of enrolling in a public plan (in my vision, modeled roughly after Medicare) through some kind of new national insurance exchange (or at least I think it should be national). There would also be private plans in the exchange that people could choose. That’s why I call the idea “public plan choice.”

My proposal for public plan choice, “Health Care for America,” is pretty much the farthest you can go with the idea. I have no special treatment of small businesses. They have the same deal as other employers: cover your workers or pay a payroll-based contribution (6% of payroll, to be exact) to cover them through the exchange—aka “play or pay.” I would fold Medicaid and CHIP into the new play-or-pay system, so people who receive benefits from these programs would either get signed up for employment-based insurance or go into the exchange like other folks without workplace coverage. Obama’s campaign proposal didn’t fold in Medicaid and CHIP, and it exempted small employers from the contribution requirement (without, smartly, saying what counted as a “small employer”). Finally, my proposal has an individual requirement that people obtain coverage (which, because of the play-or-pay requirement, would only have really mattered for those without direct or family ties to the workforce), while Obama during the campaign said he would only require coverage for kids, though he indicated he’d be open to a broader mandate down the road.

Put simply, Health Care for America is a pure, undiluted proposal for public plan choice. So it’s notable that under my proposal, according to the Lewin Group’s analysis of a couple years ago, the public plan ends up with much lower enrollment in the public plan than projected in the Lewin Group’s new analysis (90 million versus over 131 million). (For aficionados, the national insurance exchange in my proposal would have larger enrollment, but around 38 million people in the exchange would choose private plans instead of the public plan.)

Even more important, if I had kept Medicaid and CHIP separate from the exchange, as Obama indicated he wanted to do during the campaign, then enrollment in the public plan (again, according to the Lewin Group’s analysis of a couple years ago) would have been well less than half the 131 million number that the Lewin Group is talking about now: just over 56 million. Since my proposal even without including Medicaid and CHIP is still more far-reaching than the Obama or Baucus proposals, 56 million should serve as the very upper-end estimate of how large the public plan would be under the proposals embraced by leading Democrats. Quite obviously, it is much, much smaller than 131 million.

So what’s going on?

The first clue as to what is going on is that the Lewin Group provides wildly different projections for enrollment in the public plan based on how generously the public plan is paying providers. If the public plan pays Medicare rates, which are significantly lower than private rates, then enrollment in the public plan is huge. If it pays private rates, then enrollment is tiny.

This might seem obvious. If rates are lower, the premium for the public plan will be lower and more employers will enroll workers in the public plan. But that’s not how play-or pay works. Yes, the price of the public plan affects whether folks in the exchange choose it. But the really crucial question for knowing how big the public plan will be is how many people are in the exchange in the first place. And in a play-or-pay proposal, that’s determined not by what coverage in the exchange costs, but by what employers have to “pay” if they don’t “play”—that is, what payroll-contribution rate they have pay to enroll their workers in the exchange if they decide not to offer coverage.

Assuming employers are indifferent between offering coverage on their own or paying the payroll-based contribution (which is what most analyses of play-or-pay proposals, including the Lewin Group’s previous analysis of mine, assume), then employers will put their workers in the exchange when the payroll-based contribution is less than it would have cost them to provide coverage on their own. The premium for the public plan matters not at all here, and neither, therefore, does the level of the public plan’s payments for medical services.

(Yes, yes, in a larger sense, the payroll-based contribution rate could be affected indirectly by what the public plan’s premium turned out to be, because if insurance through the public plan was cheaper, you could finance expanded coverage with a lower tax rate—or lower payments from other sources, since nobody thinks the payroll-based contribution will or should cover a huge amount of the total cost of coverage. Or you could offer better benefits than you originally envisioned. Whatever. But when you are designing legislation, or modeling a proposal, you have to say what the rate is. And it’s the rate, not the premium of the public plan, that will affect the decision of employers subject to the play-or-pay requirement.)

Well, you might say, perhaps the Lewin Group just assumed that the play-or-pay requirement would apply only to larger firms, which is what Obama suggested should be done during the campaign (again, without saying what a “small firm” is). This in fact seems to be what the Lewin Group is assuming: On page 2, they say “we assume the following” and then “[l]arge employers are required to offer insurance or pay a payroll tax.” (By “large,” The Lewin Group means 10 or more workers.) In the Lewin Group’s analysis, then, smaller firms, rather than have to choose between paying or playing, would be able to voluntarily enroll their workers directly into the exchange. They don’t have to play but they can pay to cover their workers.

Now I had always though that if this were the policy (“no need to play but yes, you can pay”), then small firms would buy into the exchange on the same terms that large firms did—that is, by paying the payroll-based contribution. But the new Lewin Group analysis assumes instead that small firms would just have to pay the premium for the public plan to enroll their workers in it. Allowing small firms to buy into the public plan by paying the premium might be a more generous policy than requiring that they pay a payroll-based contribution, depending on what the contribution rate and wages of the firms were, and so perhaps this could increase the number of small firms going into the exchange, especially if you gave them a lot of subsidies for covering their workers (The Lewin Group assumes that tax credits would be provided to small employers with low-wage workers.)

But if we’re focusing just on small firms and individuals without ties to the workforce, we’re still talking about a pretty small part of that big 131 million number. According to the Lewin Group “if eligibility is limited to only small employers, individuals, and the self-employed, public plan enrollment would reach 42.9 million.” Put another way, roughly 90 million of the folks that the Lewin Group thinks are going into the public plan are working for large (in their definition of “large” as 10 or more workers) employers.

And that’s when I found the feature of the Lewin Group report that is really baffling. The payment levels paid by the public plan matter for large firms too. Check out Figure 4 in the report (reproduced below). If you look at the row that says “enrollment in national public plan” you will see that the public plan enrollment rises across the first three columns as the payment rates of the public plan go down to Medicare payment levels. These first three columns are for a proposal that only includes small firms, the self-employed, and individuals—the proposal that the Lewin Group says would result in 42.9 million in the public plan. As I’ve said, if the convoluted scenario I have described above for the treatment of small firms were really what was put in place (hey, small firm, you can buy bargain-basement coverage from us), then I could see something like this happening. Lower payments mean lower public plan premium, which means more small employers enrolling their workers in the public plan (and more individuals and the self-employed coming in); ergo, bigger public plan.


But now move to the right along the row and look at the next three columns. They show the change in public plan enrollment as rates come down to Medicare levels (lowering the cost of the public plan) under a proposal that applies to all firms. In this proposal, the large firms that are subject to the play-or-pay requirement are brought in. So we should definitely see more people go into the public plan under this scenario. What we should not see is what the columns show, that the number of people in the public plan rises as the payment levels of the public plan (and hence its premiums) fall. This is exactly what I just said cannot and does not happen under a play-or-pay proposal. The cost of the public plan has no direct bearing on the decision of employers who are subject to the play-or-pay requirement. They only care what the “pay” option entails, which is determined by the payroll contribution rate—not, I repeat not, by the public plan’s premium.

Puzzled, I wrote the Lewin Group. I received a quick and gracious reply that cleared up what was going on, but left me equally puzzled about why the Group came up with the very odd hypothetical proposal that they did.

Here are the key details of the proposal that the Lewin Group analyzed:

1. There is no exchange with private plans alongside a public plan. Employers and individuals simply buy into the public plan.

2. To buy into the public plan, all employers and individuals just pay the premium for the public plan (with subsidies for small low-wage firms and low-income individuals).

3. Employment-based health insurance would be community rated, meaning all firms would pay the same rate for the same benefit, regardless of the health status of their workers.

4. If a large firm (10 or more workers) didn’t provide coverage, it would pay a fine of 6% of payroll. However, this fine would not be applied to the cost of coverage.

So the Lewin Group’s hypothetical proposal assumes that all employers can buy into the public plan by paying its premium. There is no exchange—a key feature of Obama’s campaign proposal, my plan, Senator Baucus’s plan, and just about every other leading reform proposal. Odder still, the Lewin Group proposal says that larger employers that don’t provide coverage have to pay a 6% tax but that the tax is simply a fine, not a payment for coverage.

In other words, Lewin Group completely ignored the play-or-pay requirement. They assumed not just that small employers could buy into the public plan directly at the public plan’s premium, which is a strange assumption since it ignores the exchange and private plans within the exchange altogether. When they turned to the hypothetical proposal that gave them the headline-popping 131-million figure, they assumed that all firms—even large firms—could buy into the public plan directly simply by paying the public plan’s premium. Meanwhile, they treated the play-or-pay requirement not, in the proper way, as a central financing mechanism of reform that involves employers that don’t provide coverage paying a payroll-based contribution to cover their workers. Instead, they treated it, in essence, as a fancy penalty system for large employers.

Finally, and equally strange, the Lewin Group assumed that employment-based insurance would be subject to a rule known as “community rating”—that is, all employment-based plans would have to charge the same rate for the same benefits, without variation based on the health status of enrollees. A firm with healthy workers would pay the same amount for private coverage as a firm with less healthy workers. This, of course, raises the cost of private coverage for large firms that have healthy workforces, and thus (once again) increases enrollment in the public plan.

But that’s not an idea that Obama or any other reformer that supports the broad play-or-pay reform framework has embraced. Premiums would be community-rated in the exchange and in the individual market, but not in the employment-based insurance market.

In short, the Lewin Group’s hypothetical proposal has three features that guarantee the public plan will be big. Whatever “131 million enrollees in the public plan” represents, it does not represent an accurate picture of Obama’s campaign pronouncements, or Max Baucus’s “White Paper,” or my plan, or any other proposal under serious consideration today.

To get an estimate of how many people will end up in the public plan, therefore, we need to go back to the Lewin Group’s analyses of proposals more or less like what Obama proposed during the campaign—namely, play-or-pay plans with an exchange that features a competing public plan, and without extensive new rules on employment-based insurance. Those analyses (of my proposal or the Commonwealth Fund proposal) show that enrollment in the public plan is likely to be much, much smaller than the Lewin Group’s analysis estimates. Indeed, if you look at the Lewin Group’s analysis of my proposal (again, the maximalist public plan choice approach), more Americans have private insurance after reform than do before—either through their employer or through the new national insurance exchange.

What’s more, these are probably upper-bound estimates of the size of the public plan, since the Lewin Group assumes that employers are indifferent between providing coverage directly and paying into a public program. If, as I suspect, employers prefer to provide benefits directly and gaining the goodwill of their workers by doing so, rather than paying a tax to let government sponsor those benefits, then enrollment in the public plan is going to be even lower. (How much credit do employers get for their contribution to Social Security, after all?)

In sum, the Lewin Group’s report confirms what we knew already: public plan choice will save big money. Unfortunately, however, the Lewin Group’s report doesn’t cast any new light on the issue that has gotten the report the most attention: how many people will be in the public plan. Public plan choice will help make coverage more affordable and slow the increase of health costs. But it’s not going to massively displace employment-based health insurance—at least if we adopt the real policies advocated by leading reformers, rather than strange hypothetical policy outlined by the Lewin Group.

There’s just one final puzzle: Why did the Lewin Group analyze a proposal that’s not on the political agenda? (The proposal closest to the one the Lewin Group tried to analyze is Pete Stark’s modified “Medicare-for-All” proposal, which would allow employers to buy into Medicare.) I don’t know, but I find the obvious conspiratorial responses unlikely. No, I think the reason is probably more conventional than conspiratorial and therefore worrisome in its own way. I think the Lewin Group wanted a big media hit, and they certainly got one. But what their success shows is that on health care, the media just doesn’t have the intellectual wherewithal to dig deep into statistical models of health policy. Instead, journalists are likely to report the “top lines” without more than a passing interest in where they come from. We’re going to be seeing a lot more of this in the coming weeks. Let’s hope that our political leaders won’t lose sight of the necessity of health care reform in the fog of biased and badly done analyses.