Last night, I put up a post suggesting that the employer mandate may simply be too technically difficult to ever implement, at least as envisioned in the Patient Protection and Affordable Care Act. As long as employers are allowed to set their own eligibility rules, verification will be time-consuming and difficult to automate–not impossible, technically, but difficult and intrusive and expensive, which means that it might be impossible, politically and financially.
This has, I pointed out, big fiscal implications: offering subsidies only to people who can’t get insurance through their employer saves the government a lot of money. Allowing people to apply for subsidies under the honor system, rather than actually verifying, probably means a lot of extra money going out the door–money which, I pointed out, turns out to be very difficult to get back if you realize later that you’ve paid too much. Under the new law, the IRS isn’t allowed to claw back overpayments or collect fines the way it normally does–slap a garnishment order on your paycheck, levy your bank accounts, and in extremis, seize your home or other assets. All they can do is take it out of any payments that they owe you.
But Yuval Levin, an ultrawonk who appears to have had the entire text of the PPACA tattooed on the inside of his eyelids for quick reference, emails to say that the government’s ability to recover overpayments is even more limited than I realize:
One thought on your point that “One suspects that income verification may be not far behind; they’ll let the IRS sort it out at tax time.” The Obamacare statute really limits the ability of the IRS to sort it out and recover overpayments at tax time. And this is not only because it can only recover payments by reducing people’s income-tax refunds (which not all people have of course). There’s a more explicit barrier: In section 1401, the statute limits the amount of excess tax credit that any person with an income under 400% of poverty would have to pay back. (That begins at the very bottom of page 116 and into 117 in this final text of the law, under the section “Excess Advance Payments). The original statute limited the amount that could be clawed back to just $400. Then in the Medicare extenders bill they passed at the end of 2010 (see the table at the very end of the statute), Congress increased that amount and made it a graduated amount based on income, so it now ranges from $600 to a maximum of $3,500 for a family (half that for an individual).
CBO projects that the AVERAGE subsidy in the exchange would be worth $5,290. So the amount they’re able to claw back from people who have incomes below 400% of poverty but receive subsidies they shouldn’t (because they report a lower income than they have, falsely claim not to have been offered qualifying coverage by an employer, or report a higher income than they actually have in order to receive subsidies instead of Medicaid coverage) is likely in most cases to be significantly lower than the amount of excess payments.
“Delaying” employer reporting and income verification means more people are likely to do this and the IRS is less likely to know about it (they won’t know about people falsely claiming they don’t have an employer offer, for instance), so that even if the IRS collects back everything it possibly can at tax time, which is unlikely, there would be a major gap, and the risks people take by filing fraudulent applications are fairly limited (as you have noted before, the tools permitted to the IRS to go after excess payments are very limited).
The potential for massively expensive fraud, or even massively expensive confusion, is just enormous. The system doesn’t make sense without some meaningful prior verification, but as you say it’s not clear that such verification is technically achievable.
The problem is that the law depends on it. I’ve frequently compared the law to a Rube Goldberg machine–all the parts have to work just so for the thing to come off–and this is a perfect example. In the comments to one of my healthcare posts, commenter BronxCobra is making perfectly reasonable suggestions for a corporate IT implementation–scale back the scope, don’t do the verification real-time and instead run daily or weekly or monthly reports to verify peoples’ income against the income they reported on the form.
But this is not a corporate IT implementation where you can patiently explain that we need to reduce the scope. There’s a law. With a lot of moving parts. And those moving parts were designed for real time verification.
In part that’s for political reasons. People getting subsidies in Massachusetts right now get Commonwealth Care, a limited set of managed care plans heavily overseen by the state and apparently partially funded by a federal Medicaid waiver. You apply for it the way you do for Medicaid–send off an application, and in a few weeks you hear back. As I understand it, they do manual verifications of eligibility–not a terrible process, because the uninsured rate in Massachusetts was low even before Romneycare passed.
For obvious reasons, the Obama administration did not want to tell folks that if they needed subsidies, they would be given access to a handful of HMOs. Nor that they would go onto the exchanges and be told to print out a paper application they could send to their state’s Medicaid office. That law would never have passed Congress, or the public. So they announced that everyone above 133% of the federal poverty line would get access to the same set of plans on a nifty exchange, with the government subsidizing the premiums to make it affordable for families making up to 400% of the FPL. The exchange would automatically calculate your subsidy based on the information you gave it, and enable you to purchase insurance at the subsidized price immediately. And the exchange would verify that you actually had the income you said you did, as well as a few other things, like whether you were a citizen or legal resident.
All of this stuff had to be done up front, because one of the things they had to do in order to sell the bill was promise that the IRS would not be emptying out peoples’ bank accounts in order to reclaim excessive subsidies. Maybe that was a bad promise that shouldn’t have been made, but they did make it, and I’d argue that they had to in order to get the law passed. At this point, the IRS’ recovery abilities are so crippled that it would be folly to issue subsidies first and ask questions later.
And the subsidies have to be given up front; Folks making $50,000 a year cannot be mandated to buy a $1,000 a month insurance policy and told they’ll get their subsidy back at tax refund time.
But it’s starting to look like both of these things cannot happen, at least not in the near term. The alternatives are to delay the whole bill, or resign ourselves to hemorrhaging wads of cash. The IT expert’s instinct to hold things together with some inelegant intermediate kludge won’t work. All the elements of the law are so tightly coupled that pulling one out makes the whole machine go haywire.
Obviously, the preference of the law’s supporters is to hemorrhage cash. Just go ahead and hand out subsidies indiscriminately, the better to build political support to block repeal. But this seems . . . well, I’m struggling for kinder words, but I can’t find any. It seems wildly irresponsible. Not to mention a fundamental betrayal of the promises that were made to get the law passed in the first place.