CHIPping away at the deficit

Let’s imagine that there is a situation where a popular program can be extended.

Let’s imagine that popular program uses public-private partnerships.

Lets imagine that popular program’s extension needs no hard to agree upon pay-fors.

Let’s imagine that popular program’s extension is actually deficit reducing.

One would think that the vote to extend a deficit reducing popular program would be 422-7 in the House and 95-2 in the Senate.

That would be a nice strand of the multiverse to live in.

It is not our strand of the multiverse.

The Congressional Budget Office (CBO) estimates that extending CHIP saves money compared to the counterfactual of doing nothing.

The CBO thinks the following in their estimate:

  • Some kids currently on CHIP will get insured by going to the Exchanges
  • Some parents of kids currently on CHIP who are currently uninsured will get covered on the Exchanges
  • The repeal of the individual mandate increased projected premium subsidies
  • Fewer net kids will be covered even with spill-over coverage into the Exchanges

That to me sounds like a reasonable set of assumptions.  CHIP is cheaper for the federal government than paying Exchange subsidies for low to middle-income kids because CHIP is primarily paying near Medicare rates to providers instead of usually more than Medicare rates on the Exchanges.

A clean extension of CHIP saves money against the counterfactual of no change in policy or law.   That money could be used to enhance state match rates.  It could be used to reduce the total debt load.  It could be used for reinsurance.  It could be used for 5,001 things.

And yet, CHIP is still not being extended, 103 days after its long term funding was not renewed.



400% FPL is SOL

#400FPLSOL is my one of my unique Twitter hashtags.

Nick Bagley riffs on this at the Incidental Economist:

Because the ACA caps the premiums for those making four times the poverty level, the costs of the Trump administration’s sabotage—the repeal of the individual mandate, the impending expansion of association and short-term health plans, cuts to the outreach budget—will fall hardest on the relatively affluent, who aren’t subject to the same caps….

Consider a family of four in Arizona earning $98,000. They’re just under the 400% threshold, so they pay $9,506 in premiums for a standard silver plan (that’s 9.7% of $98,000). The federal government picks up the rest of the tab. A family earning $100,000, in contrast, has to pay full freight, or $18,348. That $2,000 increase in wages thus translates to a decrease of $8,842 in household income, or an effective tax rate of about 442%.

In practical terms, a family of four that buys private coverage in Arizona will be indifferent between earning $90,000 and $100,000….

Families will engage in financial engineering as they are better off claiming $97,000 in subsidy countable income while claiming advanced premium tax credits than earning $100,000.

We know that people were already engaged in financial engineering to qualify for ACA subsidies in the past when the stakes were not as high as they will be.

The first is an analysis of tax documentation from 2016 for people at the top of the subsidy scale:

examine the extent to which taxpayers’ incomes bunched below 400% FPL, we analyze a sample of taxpayers from 2013-2014 drawn from data maintained by the Internal Revenue Service’s (IRS) population of U.S. individual income tax returns. We find a significant amount of bunching below 400% FPL in 2014 that is only apparent among those who purchased insurance from a marketplace, and so would be potentially eligible for a subsidy. Further, we do not find such bunching among this same group of taxpayers in 2013, when the subsidies were not available. We see evidence of bunching at the cliff both among the self-employed and among wage and salary workers. We also calculate potential subsidies at 399% of FPL, and find that those whose income fell right below 400% FPL tended to be eligible for larger subsidies than those whose income fell above the cliff.
Finally, we examine the manner in which taxpayers lowered their income to keep it below the cliff, and examine whether such a response is consistent with a change in real economic activity, tax avoidance, or evasion. We see some weak evidence of reduction in wages on the left side of the cliff which is suggestive of labor response. However, the bunching appears to primarily have been driven by increases in contributions to IRA accounts.

People who make just over 400% FPL engaged in financial engineering to make their Modified Adjusted Gross Income (MAGI) shrink. This was more common in regions with high premiums than low premiums and I am betting that it was more common among older rather than younger tax filers given the nature of the age band and subsidies.

There are a lot of tools for people who have a cash flow above 400% FPL to reduce their legally reportable Modified Adjusted Gross Income (MAGI) to 400% FPL. Andrew Sprung has a good summary of the different legal ways to hide income from the subsidy calculation. Anything that moves money off the top line of the tax return seems like it works.

Families and individuals who are hovering just around the subsidy cut-off point can also restructure their employment relationships and/or reduce hours and compensation if they need to get under the line.

All of those responses are individually rational responses to implicit marginal tax rates that are well above 100% of income. All of thees responses that occur solely due the need to qualify for insurance are economically distortionary and inefficient. Removing the income cap and allowing subsidies to phase out as a function solely of income would remove significant distortions to individual behavior, but we’re not going to see that. The accountants and lawyers will be busy instead.

Maryland’s mandate plan

Maryland has a really interesting plan to use a state based individual mandate and a kissing cousin of auto-enrollment to maintain their individual insurance markets. They call is the Downpayment Plan.

The plan starts in 2020. An individual mandate will be assessed. Individuals paying the mandate will be notified if they qualify for advanced premium tax credits (APTC). If the combination of APTC and individual mandate penalty makes the monthly net of subsidy premium be equal to zero, the person is auto-enrolled in a plan. If the cost of the least expensive premium is more than the APTC and individual mandate, the individual mandate collection is held in escrow for a year to help pay for insurance in the next open enrollment. If there is no active selection in that second open enrollment period, the held in escrow individual mandate payment is transferred to a state insurance stabilization fund where it is presumably used for reinsurance or subsidies for individuals who don’t qualify for federal APTC.

My first reaction to this is that it is nifty and creative. It also highlights the extreme option value of a state running their own exchange. I don’t know if could mechanically do what Maryland wants to do.

Secondly, the program will wildly vary across age and county. In Baltimore City, a 41 year old earning $30,000 a year qualifies for a $0 net of advanced premium tax credit and individual mandate Bronze plan. However in Alleghany County (Western Maryland) a 21 year old earning $30,000 qualifies for a Gold plan under this same scheme. The difference is due to regional Silver levels and Gold loading plan offerings. Baltimore has a pair of fairly inexpensive and tightly clustered Silver plans offered by Kaiser. Alleghany County has a cheap Silver HMO offered by CareFirst and then an ungodly expensive Silver PPO also offered by CareFirst that acts as the benchmark.

In this system, inherently lower premiums for the Silver benchmark is not necessarily a good thing. Very active plan management by the state in order to maximize the Silver on-Exchange benchmark while also minimizing on-Exchange Bronze premiums and off-Exchange Silver premiums would optimally be needed. A hyper narrow network provider that offers multiple low premium Silver plans dramatically reduces the number of people who can qualify for a zero-premium after APTC and individual mandate plans.

Overall, this is interesting and it can be replicated in other states that run their own marketplaces and are willing to actively manage plan offerings.

State based mandates

Nick Bagley, a law professor at the University of Michigan, also has a model state individual mandate law.

State based mandates, if adapted, could give states significantly more flexibility in future plans for 1332 waivers in 2019 and beyond. Under current law, mandate revenue goes to the federal general fund to offset some of the costs of providing subsidies and expanding Medicaid. None of that money goes to the state.

Now let’s imagine a world where there is no federal mandate but a state mandate that mimics the current federal mandate. What happens?

Some people will get insurance who otherwise would not have as the relative costs of being uninsured increase compared to buying minimal coverage.

Some people will still not get covered and will not be exempt. The state will collect the individual mandate from these people.

And now a state has a pool of money. It is not a huge pool of money, the IRS collected about $3 billion nationwide for tax year 2016. But it is not nothing.

This would be unanticipated revenue. One of the ways that might make state based individual mandates more palatable at the state level would be if the revenue was dedicated to bringing down premiums for non-subsidized buyers. These funds could be used for reinsurance that will bring down index premiums for subsidized and non-subsidized buyers which mainly benefits non-subsidized buyers. It could be used for direct premium subsidies for people who earn more than 400% ($48,080) of the Federal Poverty Line (FPL). It could be used to start-up and back a state public option that only sells off-Exchange policies.

Furthermore, if a state was to enact an individual mandate, it will lead to a higher enrollment levle and a lower Silver benchmark. This will lead to a different set of calculations for any 1332 waivers that they may wish to submit. The per member per month (PMPM) pass through amount may decrease but the PMPM will be multiplied against a larger number of covered lives. The overall sum of money available for a 1332 may increase compared to the counterfactual of no state based individual mandate. Larger potential 1332 pass-throughs means more state flexibility if a state wishes to engage in customizing their healthcare system.

Update 1 The downside of pre-writing posts is that Maryland puts out an interesting proposal to use a state level mandate and auto-enrollment after you write and before you publish.

The ACA in 2018 and Obama’s 2007 vision

This is worth at least an quasi-ironic chuckle.  The implementation of the Affordable Care Act in 2018 is closer to some elements of the early 2007 campaign version of Obama’s health care vision than the ACA implementation from 2010-2017 was.

Now let’s go look at the tape, from Polifact 2007:

Obama shot back: “Well, let’s talk about health care right now because the fact of the matter is that I do provide universal health care. The only difference between Sen. Clinton’s health care plan and mine is that she thinks the problem for people without health care is that nobody has mandated — forced — them to get health care.”….

One of the few differences is that Clinton and Edwards include a universal mandate….

Obama’s decision not to include a mandate is a more cautious approach, one Obama says is designed not to penalize people with modest incomes. If premiums don’t drop enough after all the reforms are implemented, people will still be unable to afford insurance.

Obama contended during the 2008 primary that the subsidies would be rich enough that no mandate would be needed. And then the desire to hit a particular CBO score came into play as well as a need to not lose a single Democratic senator dictated that subsidies were going to be a lot weaker than the House wanted.

However with Cost Sharing Reduction (CSR) sabotage leading to Silverloading and Gold gapping, subsidies for individuals who make between 100% Federal Poverty Line (FPL) and 400% FPL ($12,020-$48,080) will have become effectively much richer for people who buy plans that cost less than the benchmark.

This is just an irony if seen from a skewed point of view.

There will be lawyers (CSR Edition)

Health Affair’s new ACA blogger, Katie Keith, writes about the first of many lawsuits concerning Cost Sharing Reduction (CSR) subsidies:

On December 28, 2017, Maine Community Health Options (MCHO)—a nonprofit insurer in Maine—filed what is believed to be the first lawsuitagainst the U.S. Department of Health and Human Services (HHS) for failing to reimburse marketplace insurers for cost-sharing reductions (CSRs) for 2017. MCHO seeks an estimated $5.6 million in CSR payments for the 2017 plan year….

MCHO’s claims are relatively straightforward. In brief, MCHO alleges as follows. As an insurer participating in the marketplace, it is required to offer CSR plans and is guaranteed to be reimbursed by the government for doing so under Section 1402 of the ACA and its implementing regulations at 45 C.F.R. 156.430. By failing to make CSR payments, the government deprives insurers of funds that they are statutorily entitled to for participating in the marketplace in 2017. The lack of congressional appropriation is irrelevant because the government has a statutory obligation to make CSR payments and insurers have the right to receive them.

The language in the law is that the insurers “SHALL” offer CSR and the government “SHALL” pay insurers for this added benefit on a regular and timely basis.  I am not a lawyer but I know “SHALL” is a very powerful word denoting strong obligations.

There are two sets of CSR lawsuits that are possible.  The Maine Co-op is filing the first type that should be the most straightforward.  They are asking only for CSR payments for the last three months of 2017 as they were able to mitigate their damages in 2018 by raising their Silver rates.  This one is fairly straightforward.

The more complex potential CSR lawsuit starts with the 2017 lawsuit and then adds in an ongoing recoupment of costs after net mitigation through Silver Loading.  That lawsuit would be complex and a guarantee only of billable hours.


Risk pool surprises and rates

My basic heuristic for analyzing the individual health insurance market under the Trump Administration so far is the following:

  • >=100% Federal Poverty Line (FPL)($12,020) to <=400% FPL ($48,080) Okay or better off
  • <100% FPL minimal changes
  • Healthy people above 400% FPL better off
  • Not health people over 400% FPL screwed

There are some serious caveats (Medicaid work requirements for instance) but this is my short hand heuristic when I look at the individual market.  The recent proposed rule for expanded association health plans and the soon to be released proposed rule on short term plan expansion fits into this heuristic.

The ACA had a significant number of channels out of the regulated individual and small group market.  Individuals could pay the mandate, they could enroll in a health sharing ministry, they could have  Tennessee Farm Bureau coverage, they could maintain Grandfathered and Grandmothered coverage.  Short duration plans are just another out.

Short duration plans are underwritten.  Their risk pools will be overwhelmingly healthier people.  The people who remain in the ACA risk pool will be sicker and more expensive after short term plans are expanded than before.

Insurers deal with bad risk pools by raising rates.  A typical constraint on raising rates is the fear that higher rates will lead to a death spiral.

Kaiser Family Foundation has been tracking premiums and claims through the third quarter of the individual market on a per member per month (PMPM) basis since before the guaranteed issued and community rating reforms were implemented on 1/1/14.


Looking at the Kaiser graph, we see the ACA discontinuity in claims.  Claims went up on a PMPM basis by 38% between Q3 2013 and Q3 2014.  Premiums did not increase fast enough on the initial attempt due to a combination of expectations that risk corridors would act as a shock absorber, winners curse on the subsidy bid structure and an assumption by some insurers that profitable ACA members would be far stickier than they have turned out to be.

Insurers responded to the new information of what the ACA risk pool actually looks like by raising premiums.  As we’ve chronicled here, the base premium does not matter much for subsidized buyers.  It matters a lot for non-subsidized buyers.  Premiums were probably raised a bit too high for 2017 compared to the underlying claims.

Insurers will respond to a deterioration of the ACA community rated/guaranteed issued risk pool by raising premiums.  And this won’t matter to subsidized buyers.  Healthy subsidized buyers will leave the market and the people who make too much for a subsidy but can’t get underwritten are whacked.