• How reasonable is the White House’s marketplace enrollment goal?

    This is a guest post from Adrianna McIntyre, a graduate student studying health policy at the University of Michigan. Adrianna currently blogs at Project Millennial, tweets at @onceuponA, and will be joining TIE this fall as the team’s first intern. The analysis was conducted with generous assistance from Josh Fangmeier, a health policy analyst based in Ann Arbor.

    There’s been lots of fuss about young adults and Obamacare, from the bro-tastic “rate shock” debate to more recent coverage of White House outreach efforts (you should read Ezra Klein and Sarah Kliff on that). As a bona fide twenty-something, you could say I have a vested interest in the matter.

    One figure that’s been bandied about is 2.7 million—the alleged magic number of 18-35 year olds required to keep premiums stable in the new marketplaces (the new term for “exchanges”). That number comes from the White House, though I’m not sure how they arrived at it or how important it really is. But let’s accept it for the sake of argument. Is it a reasonable goal?

    It’s worth understanding how the target population—young adults who are uninsured or currently have nongroup insurance—breaks down. About 25 million people between the ages of 18 and 35 match that description, stratified below along age and income bands important to the health law.

    young adults-Figure1

    The red bars represent the number of young adults living below the poverty line in states opposed to expanding Medicaid—they won’t receive Medicaid coverage, nor will they qualify for marketplace subsidies—while the orange shows how many adults in expansion states are expected to enroll in the program. These numbers will include “woodwork” individuals who are currently eligible for Medicaid but have not signed up for coverage.

    The dark blue bars capture those between 100-200% FPL who qualify for subsidies (including those between 100-138% FPL in states not expanding Medicaid). Green and purple bars capture those with incomes too high for Medicaid eligibility even in expansion states, but not too high to qualify for marketplace subsidies. Finally, the light blue bars capture those with incomes above the subsidy cutoff of 400% FPL.

    Age matters, too. Individuals under 26 can stay on their parents’ insurance, if the plan includes coverage for dependents. Those under 30 have the option to buy catastrophic plans, though these plans are not eligible for subsidies like “metal-tier” plans—and recent insurance filings suggest that their premiums are quite similar to basic bronze plan alternatives. Catastrophic plans have separate risk pools, so they don’t count toward marketplace enrollment targets.

    If we look at just the predicted marketplace target population—setting aside those who will receive coverage under the Medicaid expansion and those below the poverty line in nonexpansion states—we’re left with approximately 11 million adults between the ages of 18 and 35. The White House target is about a quarter of that, though this doesn’t account for dependent coverage (about 4 million are under 26) or take-up of catastrophic plans.* Even so, this strikes me as achievable, especially since their incomes are predominantly below 300% FPL, meaning most will receive subsidies to offset the cost of premiums.

    young adults-Figure2

    There are caveats to the Administration’s goal. Notably, marketplaces are state-based, so even wildly successful outreach to young adults in Connecticut and Oregon won’t do a thing for my premium in Michigan. Additionally, the messaging about how we “need” 2.7 million young adults uses “young” as a proxy for “healthy”. Enrolling healthy adults, across all age groups, is what would really keep premiums down in the marketplaces.

    I do have some lingering concerns about the scaling-up of marketplaces—the CBO estimated that enrollment would grow from 7 million in 2014 to 24 million in 2016. If actuarial estimates suggest that we need to maintain a certain ratio of young-to-old, the fact that young adults disproportionately benefit from the Medicaid expansion and extended dependent coverage could put stress on the marketplaces.  Future decisions to expand Medicaid will shift adults between 100-138% FPL from marketplaces into the public program; in states that currently oppose expansion, there are about 1.5 million 18-35 year olds in that income band.

    I’m optimistic that the White House will reach their 2.7 million target. Even if I’m wrong, I think it’s an overreach to suggest that success of the entire health law hinges on this figure. Marketplace premiums have risk-selection assumptions built in—what matters is that actual enrollment matches those assumptions, not the Administration’s goal.

    Note to ultrawonks: You can access a description of the methodology used to compile relevant Census data here, and a spreadsheet of the data itself here.  

    * UPDATE: The provision extending dependent coverage up to age 26 went into effect in September 2010, before this ACS data was collected. It is possible that more uninsured or privately insured individuals have taken advantage of that option since this was collected, which would reduce the number of 18-25 year olds likely to seek coverage on the exchanges. Additionally, it’s possible that some 18-25 year olds in the nongroup market are dependents on a parent’s nongroup plan. The under-26 requirement only extends to plans that already offer dependent coverage.

    • This is an admirable and comprehensive attempt to analyze a very complex problem. Unfortunately, it only serves to highlight the many very grey areas and assumptions that have been made with too little information.
      One could start a discussion of these grey areas and assumptions but I fear this would just add to the complexity without shedding any light on the problem. There are just too many “unknown unknowns” (and perhaps unknowable unknowns).
      I think we will just have to wait until the actual numbers arrive and then start a useful discussion on what could be done to improve them if they are low or congratulate ourselves if they are high.
      Until then, I fear our discussion would be primarily a reflection of our biases… what we would want to happen.

    • I agree with Mark. Those are some bold assumptions being made. Noting the educational-income relationship, what leads you to believe that the 18-30 year old individuals targeted will be made aware of the Exchanges. understand how to enroll, and understand how they operate? Take the recent Oregon Medicaid Experiment paper as an example. Finkelstein et al. (2012) reported that 40% of the newly Medicaid-eligible individuals did not return their applications. Why was that? It seems that the Obama Administration is concerned about this too. Why else would they be reaching out to the NFL, Hollywood, etc. in what seems like a desperate, last-minute attempt?

      Similarly, what leads you to believe that the 18-30 year old individuals will want to enroll, even if they are fully aware of the Exchanges? The fines for the first year are still cheaper than paying for annual coverage. Most will argue that the subsidies may reduce costs substantially, leading these individuals to enroll. Regardless of that, if they don’t want to, why would they? As an example, the AHRQ reported that, in 2011, 23.9% of employees offered employer-based insurance did not enroll; although, there could be a number of reasons for this. The AHRQ did report that this number is at an all-time low, though.

      On another note, I would also like to know more about these individuals. Why are they living under 400% of the FPL? Are they college students? Are they recent graduates struggling to find employment? None of the blogs or papers I have been reading seem to have these answers. That data could effect the analyses of who and how many will actually enroll. By all means, though, if someone has information on this, please pass it along.

      Sources (please forgive the poor citation format):

      AHRQ: http://meps.ahrq.gov/data_files/publications/st389/stat389.pdf

      Finkelstein et al. (2012):

    • I’ve offered what Census data can inform—I agree that it’s far from perfect, and I try to get at some of the important limitations in the attached methodology. Regardless, I think any conversation about enrollment of young adults ought to have some context provided about how large the relevant population is, and how it’s distributed across the income spectrum. So far, that hasn’t been the case.

      To your point about Finkelstein et al, the paper also reports that about half of individuals who applied for Medicaid coverage after being selected in the lottery didn’t qualify, mostly due to having incomes above the maximum threshold. It’s likely that at least as many who didn’t apply were similarly ineligible. Some (such as Jim Manzi) have posited that low take-up in OHIE is a signal of “low prudence.” To the extent that this is true, a longer enrollment period may mitigate some of that effect (OHIE participants had a month to enroll in Medicaid; initial exchange enrollment will run from October 1, 2013 through March 31, 2014).

      There’s evidence from survey data that young adults are interested in having health coverage. Your point about take-up of ESI is an important one, and something I acknowledge in the methods. I also can’t speak to the college student question, though that is likely to have a lot of interaction with the under-26 provision. If you wanted to eliminate that cohort from consideration entirely, you’re still left with about 7 million 26-35 year olds above the poverty line who were uninsured or had nongroup insurance when this data was collected (2011).

      • Adrianna,

        Excellent analysis and very helpful – as far as it goes…

        And it is not any fault of yours that it can’t go further…

        As Mark notes, we face a lot of “unknowables” going forward…

        I am a bit surprised that we have not seen a good study on “intent to participate” – perhaps we will soon…

        I am also surprised that we don’t have a good study on “expectations”.

        One of my fears is that many who sign up for a plan will find the plan “disappointing” – either when they read the fine print or when they first make use of it. The annual out of pocket amounts will surprise (and disappoint) many.

        For those closer to the FPL they may not have the financial reserves to pay several thousand in OPE.

        These known “unknowables” are bad enough – wait until we starting to surface unknowable “unknowables” as we recently did with the delay in the employer mandate. We can only guess at what the delay combined with an honor system will do to the amount of both intentional and accidental subsidy overpayments.

        We can probably agree that the impact of is a higher cost for the program in 2014 than it would have been with the mandate in place – we just don’t know how much higher it will be.

    • When we get closer to open enrollment, will IE sponsor a pool so we can bet on what enrollment will actually be?

    • As a followup to my post on unknowables I spent a good deal of time looking for things that might inform some of my thinking on these…

      And found this

      I have spent a little time with it – and intend to spend more – but will confess that so far I find myself more nervous about how this will play out than less.