John McCain took a stab at solving his health insurance dilemma Tuesday. No, not the question of how he, as a cancer survivor, might buy insurance on the open market -- as a senator married to an heiress, his health insurance is covered. I mean his policy conundrum: how to reconcile popular demands for more access to better health care with "free market" principles that would appear to make universal coverage even harder to achieve.

The result is an interesting modulation of his doctrinaire position, as elaborated in a speech at a cancer center in Tampa. No longer does he propose thoroughly dismantling the employer-driven health insurance system and replacing it with a market of individual consumers. Instead, he merely wants to weaken protections for the employer system so that the two markets can sit side-by-side. He'd do this by taxing the insurance benefit that individuals get from their employers, which is currently excluded from income tax. Then he'd give everyone a tax credit to buy insurance anywhere they can find it: $2,500 for individual coverage and $5,000 for families. (McCain's website has posted both a summary of his proposal as well as the text of his speech. But while the press release repeatedly and emphatically insists that the proposal is loaded with "specifics," I found both the speech and summary maddeningly vague, and had to rely on two accounts from the New York Times, here and here. Ultimately, I had to get further clarification from the McCain campaign.)

At first sniff, McCain's tax credit has the slight odor of Democrat about it. For one thing, most of the major Democratic presidential candidates proposed a tax credit to help people pay for health insurance. And for poor and working class people, McCain's $5,000 credit is more generous than the tax deduction Republicans (including, up to now, McCain himself) have proffered. For instance, Rudy Giuliani's (remember him?) $15,000 deduction, works out to roughly equivalent to a $5,000 credit only at the higher tax brackets, and McCain would make his credit available even to people who pay no taxes. Then there's the (perhaps unintended) irony that while most Americans under this proposal would get a tax cut, since the credit would exceed the levy on the premium benefit, those who are highly paid or have particularly generous health benefits will pay more in taxes -- a notion that warms my bleeding heart. McCain's top policy adviser, Douglas Holtz-Eakin, told the Times that any employee in the top tax bracket whose company contributed $14,285 toward insurance premiums would see a higher tax bill. According to the Times, that could account for up to six percent of insured employees.

But as I've said before, it's tricky to balance the individual market against the group market -- until now, it's always been a zero-sum game. When the young and healthy leave the group for a better deal in the private sector, what's left of the pool has to pay more. And traditionally when people (Republicans) talk about incentives to stimulate the individual market, like an individual tax deduction, it usually comes at the expense of the employer tax deduction. (Until Tuesday, McCain himself wanted to shift the income tax deduction for premiums from employers to individuals.) So McCain is attempting something novel with his balancing act, and the consequences may be novel, too.

It turns out that McCain's plan may do little to move employees at big firms into the individual market -- but it could decimate the small group market where the nation's entrepreneurs buy health insurance. Compare average costs for employer health insurance (from the Kaiser Family Foundation) against averages for the individual market (from America's Health Insurance Plans, an insurance trade association), and you'll see that on a strictly economic basis, no employee seeking coverage for just himself would willingly dump his firm's policy for something from the marketplace. The problem is the deductible. For young employees, it's not hard to find individual policies on the individual market that are cheaper, on average, than individual policies available from work, even after you calculate the employer's contribution to premiums and the taxes on those benefits. (If you want to get all wonky about it, click here.) But deductibles are so much higher in the individual market than the group market -- they just kill the savings for every age group. Even an 18-year-old will usually find insurance through work a better deal.

Family coverage is more muddled, because the premiums are nearly triple individual coverage in the group market, but don't rise nearly so high in the individual market. Moreover, workers shoulder a larger share of the family policy premium, especially at companies with fewer than 200 employees. Finally, the disparity between deductibles in the two markets is not nearly so stark in this case. The upshot is that if you're a family man (or woman) under 30 at a big firm, you may well be tempted to make the switch. At small firms, 45 is the new 30: anyone younger might find out-of-pocket costs lower in the individual market.

The question becomes: what happens to the group if all those young (and for small firms, not so young) families decamp for the private market? It's a particularly pressing problem for small groups (usually under 50 employees), where are premiums are based on the demographics of the group. (Big groups -- those that don't self-insure -- are usually pooled into a much larger aggregation of big groups, so the specific contours of any particular group doesn't really matter.) "Selection is always an issue when it comes to designing insurance markets," Holtz-Eakin said when I put the question to him. But by stimulating the individual market, "you also change the kinds of coverage that those policies are going to offer. You'll have a genuine incentive for an insurer to cover the preventive care." Eventually, he said, "that'll also show up in the employer market, and that's what those older workers need -- that's where you get the big gains. So it's not obvious that you're going to get into this price spiral that people worry about." Though Holtz-Eakin was not willing to predict whether those costs will decrease in tandem between the markets, I'm willing to guess that an older group will see less of a price decrease than a market of younger individuals. And, of course, if efficiencies in the individual market outpace efficiencies in the group market, still more people might leave the group market, making it even more expensive -- the "spiral" Holtz-Eakin mentioned. On the other hand, if few employees make the switch, the individual market will be stillborn, and so, too, will the hoped-for efficiencies.

It's rare that I get to indulge by inner pundit but: only time will tell.

Here's how I arrived at my conclusions for the average working stiff at a big firm (200 or more employees) with individual coverage. According to Kaiser, the average premium is $4,442, of which our man pays $759 and his employer picks up $3,683. If his total tax rate (federal, state, and local) is 30 percent, then under McCain's plan he'd pay $1,105 in taxes on the $3,683 benefit. But he'd get a $2,500 tax credit from the government, so the net credit would equal $1,395. Since he only pays $759 in premiums, under McCain's plan he ends up $636 ahead.

If our Joe Sixpack goes comparison-shopping in the individual market, the pressing question is whether he can buy a policy there for less than $1,864 ($2,500 minus $636). If he can, he's saving money; if he can't, he might as well stay put. And it turns out there are plenty of policies out there for under $1,864 if you're young enough. According to a survey by America's Health Insurance Plans, 30-34 year-olds pay an average of $1,877 in premiums. Anyone 35 or younger will do better in the individual market.

But: the average deductible for a Preferred Provider Organization or Point of Service plan in the individual market is a whopping $1,747, while the average PPO deductible at a big firm is a more modest $382. When you add the $1,365 differential back to the average premiums in the individual market, none of them even come close. Total out-of-pocket costs for an 18-year-old would rise to $2,528, while a $30 would pay $1,342 -- nearly $1,400 more.

Obviously, these are not rigorous calculations. I didn't try, for example, to weight the PPO premiums in the group market to equalize it to the individual market -- I don't get paid enough for that kind of manipulation. But it wouldn't change the broad conclusions if I had. You get the idea.