The vote on repealing the health care bill looks likely to be delayed for a while, which gives us more time to quarrel over what such a repeal would mean.

Not what it will mean, of course--Republicans don't have the votes to get it through the Senate, so this is largely a symbolic move.  But there's still plenty of room to argue about what would happen if they managed to retake the Senate and get it done.

Democrats are pointing to the CBO score on repeal, which estimates that repealing the law would cost over $200 billion.  Republicans are pointing out that the score is highly contingent on actually implementing the law as passed--something that even the bill's proponents recognize will be a political challenge. Already the high-risk pools are a disaster, and the provision to require greater use of 1099s is generally agreed to be a bad idea; we haven't even begun to do the hard stuff, like cutting back on services in order to hold down costs.

Ruth Marcus has a smart take on this in a recent blog post:


Of all the cost estimates that the CBO produces, the most complex and least reliable involve health care. This is in no way a criticism of CBO. No matter how sophisticated the economic model, the multi-layered assumptions about the future cost of health spending make the $230 billion projection closer to an educated guess (albeit a guess made by very educated economists) than a take-it-to-the-bank certainty.
And even if the CBO forecasts were guaranteed to come true, there is the political calculus to consider. CBO made this point in its usual, restrained way Thursday, noting that "current law now includes a number of policies that might be difficult to sustain over a long period of time. If those policies or other key aspects of the original legislation would have been subsequently modified or implemented incompletely, then the budgetary effects of repealing [the health-care law]...could be quite different."

Translated into English: don't bet on that $230 billion. The health-care law will require billions in new spending. It relies on the expectation of billions in savings from slowing the growth of health-care costs and assorted cuts and taxes -- all guaranteed to produce howls of outrage, and a burst of lobbying, from the affected interests. Health-care reform, done right and with steadfastness that is not always forthcoming from the legislative branch, could be a huge contributor to reducing the deficit.

But the costs of the new law are far more certain than the savings. Anyone who's spent any time in Washington knows better than to assume that health-care reform will end up as a money saver.

Even beyond these complexities, there's the question of the short time frame of CBO projections.  Take the question of employee dumping, recently discussed by Marcus's colleague, Ezra Klein.  The fear is that the availability of subsidized insurance on the health care exchange will cause employers to "dump" their employees onto the exchanges, costing the government much more money in subsidies than we are currently expecting.  Ezra writes:


According to an informed source, the reason CBO doesn't worry about dumping is that it gets canceled out: If a company dumps its workers on the exchange, it also loses the massive tax break it gets for their health care. So the government may have to provide subsidies for some of the workers, but it also gets more tax revenue from the employer. Meanwhile, the employer has to pay those workers higher wages to compensate for taking away their benefits, and those wages aren't tax-free.
 

The bigger concern with employers, my source said, was that they'll design high-deductible plans that push sick employees onto the exchange. This would be fairly easy to stop if it ever began happening in large numbers, but if Congress refuses to do anything to fix problems with the law, then it would be harder to stop.

Let's unpack this a bit.  During health care reform, I had a similar conversation with a similarly anonymous source familiar with the CBO's health care scoring process.  I got essentially the same answer about both dumping and the excise tax: it doesn't matter what the employers do, because they expect that any cut in benefits will eventually be replaced by higher wages--wages which are taxable.  Assuming that the higher wages are essentially a dollar-for-dollar replacement for the lost benefits, then the increased tax revenue will make up for the need to subsidize some of those employees.


Unlike Ezra, my understanding was that this is essentially a question of income taxes, not taxes paid by the employer--wages are also tax deductible to the employer, and economists believe that the "employer side" of our payroll taxes ultimately comes out of wages, not profits, as employers compensate for the loss by offering lower salaries.  This is not hugely important, but it does factor into the major problem with this argument, which is that it is probably true over the short term, but not over the long term.

The minor problem with this argument is that it's conceivable that employers could find a way to transfer at least some of this compensation into other tax-advantaged forms, such as life insurance, plusher offices, etc.  This doesn't strike me as very likely, but since it is possible, it needs to be considered.

The larger problem is that the workers who are eligible for subsidies don't pay much in the way of income taxes.  The effective tax rate for the bottom two income quintiles is negative--workers get more back in refundable tax credits like the EITC, than they pay in federal taxes. Those workers are the ones who qualify for the biggest subsidies.

Ah, you will say, but they do pay payroll taxes!  Absolutely correct.  Unfortunately, the vast majority of payroll taxes are social security "contributions".  And social security contributions differ from the income tax in one crucial respect: they obligate the government to pay more benefits in the future.  An average wage-earning couple will get back about $5 out of every $6 they pay in Social Security taxes.  Lower wage earners will get even more because the formula for calculating benefits is fairly progressive--though this is complicated by the fact that lower-income workers are also disproportionately likely to die before they collect much in the way of benefits.

Now, the Social Security system may change over the years, but given that the main proposal on the table is the Simpson-Bowles plan to make the system even more progressive, it's not particularly likely that it will change in a way that on net takes in more tax revenue from the workers poor enough to qualify for a sizeable subsidy.  And the CBO scores what's written into law today, not what might happen; given CBO head Doug Elmendorf's publicly expressed skepticism about the political sustainability of some of Obamacare's provisions, this is probably the only reason it scored as deficit-reducing in the first place.

What this means is that within the CBO's scoring window of ten years, dumping won't be very costly, but over the long run, it could prove very costly indeed if most of the workers who are "dumped" are lower-wage workers who qualify for substantial subsidies.  Over the short term, the government will recoup some of its expenses through higher payroll taxes--but over the long run, workers will get most of that tax money back in the form of higher social security benefits.  Since that "long run" mostly falls outside of the CBO's scoring window, the forecast may mislead us into thinking that dumping is less costly than it will ultimately prove.

I don't mean to say that the CBO, or Ezra, are being misleading; the scoring process inevitably has limitations, and this is one of those artifacts.  But thinking through these questions make you realize just how complex, and contingent, these sorts of projections are.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.